From 44b477c3ad923a3c598ca19b45e5a4b249703e62 Mon Sep 17 00:00:00 2001 From: rafelafrance Date: Fri, 9 Feb 2024 09:02:51 -0500 Subject: [PATCH] Remove setup.py --- README.md | 12 +++---- finder/__finder_label_babel_reconcile.py | 3 +- finder/__finder_rise_of_machines_build.py | 3 +- finder/__finder_rise_of_machines_reconcile.py | 3 +- finder/yolo_inference_data.py | 3 +- finder/yolo_results_to_labels.py | 3 +- finder/yolo_training_data.py | 3 +- pyproject.toml | 10 ++++-- setup.py | 34 ------------------- tests/pylib/test_box_calc.py | 4 +-- 10 files changed, 26 insertions(+), 52 deletions(-) delete mode 100755 setup.py diff --git a/README.md b/README.md index caa0ab5..891caf1 100644 --- a/README.md +++ b/README.md @@ -12,16 +12,14 @@ We find labels with a custom trained YOLOv7 model (https://github.com/WongKinYiu - All other identified labels are outlined in teal. Local scripts: -- `build_training_data.py`: This formats images of herbarium sheets to prepare them for training a YOLO model. +- `fix-herbarium-sheet-names`: I had a problem where herbarium sheet file names were given as URLs and it confused the Pillow (PIL) module so I renamed the files to remove problem characters. You may not need this script. +- `yolo-training`: If you are training your own YOLO7 model then you may want to use this script to prepare the images of herbarium sheets for training. The herbarim images may be in all sorts of sizes, and model training requires that they're all uniformly sized. - This script also requires a CSV file containing the paths to the herbarium sheets and the class and location of the labels on that sheet. -- `build_inference_data.py`: This formats herbarium sheet images for YOLO inference. - - It requires a CSV file containing the paths to all of the herbarium sheets. -- `results_to_labels.py`: This takes for output of the YOLO model and creates label images. The label name contains information about the YOLO results. The label name format: +- `yolo-inference`: Prepare herbarium sheet image for inference; i.e. finding labels. The images must be in the same size the training data. +- `yolo-results-to-labels`: This takes for output of the YOLO model and creates label images. The label name contains information about the YOLO results. The label name format: - `_