add readmes

integration/monai/examples/README.md

@@ -1,5 +1,9 @@
 # Examples of MONAI-NVFlare Integration
 
+### [Converting MONAI Code to a Federated Learning Setting](./mednist/README.md)
+A tutorial to show how simple it can be to run an end-to-end classification pipeline with MONAI 
+and deploy it in a federated learning setting using NVFlare.
+
 ### [Simulated Federated Learning for 3D spleen CT segmentation](./spleen_ct_segmentation_sim/README.md)
 An example of using [NVIDIA FLARE](https://nvflare.readthedocs.io/en/main/index.html) 
 to train a medical image analysis model using federated averaging ([FedAvg](https://arxiv.org/abs/1602.05629))

integration/monai/examples/mednist/README.md

@@ -0,0 +1,18 @@
+## Converting MONAI Code to a Federated Learning Setting
+
+In this tutorial, we will introduce how simple it can be to run an end-to-end classification pipeline with MONAI 
+and deploy it in a federated learning setting using NVFlare.
+
+### 1. Standalone training with MONAI
+[monai_101.ipynb](./monai_101.ipynb) is based on the [MONAI 101 classification tutorial](https://github.com/Project-MONAI/tutorials/blob/main/2d_classification/monai_101.ipynb) and shows each step required in only a  few lines of code, including
+
+- Dataset download
+- Data pre-processing
+- Define a DenseNet-121 and run training
+- Check the results on test dataset
+
+### 2. Federated learning with MONAI
+[monai_101_fl.ipynb](./monai_101_fl.ipynb) shows how we can simply put the code introduced above into a Python script and convert it to running in an FL scenario using NVFlare.
+
+To achieve this, we utilize the [`FedAvg`](https://arxiv.org/abs/1602.05629) algorithm and NVFlare's [Client
+API](https://nvflare.readthedocs.io/en/main/programming_guide/execution_api_type.html#client-api).

integration/monai/examples/mednist/monai_101_fl.ipynb

@@ -34,7 +34,7 @@
     "\n",
     "# MONAI 101 tutorial with Federated Learning\n",
     "\n",
-    "In this example, the **server** uses the [`FedAvg`](https://github.com/NVIDIA/NVFlare/blob/main/nvflare/app_common/workflows/fedavg.py#L44) controller, which performs the following steps:\n",
+    "In this example, the **server** uses the [`FedAvg`](https://github.com/NVIDIA/NVFlare/blob/main/nvflare/app_common/workflows/fedavg.py) controller, which performs the following steps:\n",
     "1. Initialize the global model. This is achieved through the method `load_model()`\n",
     "  from the base class\n",
     "  [`ModelController`](https://github.com/NVIDIA/NVFlare/blob/fa4d00f76848fe4eb356dcde417c136047eeab36/nvflare/app_common/workflows/model_controller.py#L292),\n",
@@ -56,8 +56,8 @@
     "6. The last step is to save the updated global model, again through\n",
     "  the [`ModelPersistor`](https://nvflare.readthedocs.io/en/main/glossary.html#persistor).\n",
     "\n",
-    "The **clients** implement the local training logic using NVFlare's [`Client\n",
-    "API`](https://nvflare.readthedocs.io/en/main/programming_guide/execution_api_type.html#client-api)\n",
+    "The **clients** implement the local training logic using NVFlare's [Client\n",
+    "API](https://nvflare.readthedocs.io/en/main/programming_guide/execution_api_type.html#client-api)\n",
     "[here](./code/monai_mednist_train.py). The Client API\n",
     "allows the user to add minimum `nvflare`-specific codes to turn a typical\n",
     "centralized training script to a federated client-side local training\n",