Federated Learning with MONAI and Flower (Quickstart Example)

This introductory example to Flower uses MONAI, but deep knowledge of MONAI is not necessarily required to run the example. However, it will help you understand how to adapt Flower to your use case. Running this example in itself is quite easy. MONAI(Medical Open Network for AI) is a PyTorch-based, open-source framework for deep learning in healthcare imaging, part of the PyTorch Ecosystem. This example uses a subset of the MedMNIST dataset including 6 classes, as done in MONAI’s classification demo. Each client trains am DenseNet121 from MONAI.

[!NOTE] This example uses Flower Datasets to partition the MedMNIST dataset. Its a good example to show how to bring any dataset into Flower and partition it using any of the built-in partitioners (e.g. DirichletPartitioner, PathologicalPartitioner). Learn how to use partitioners in a step-by-step tutorial.

Set up the project

Clone the project

Start by cloning the example project:

git clone --depth=1 https://github.com/adap/flower.git _tmp \
        && mv _tmp/examples/quickstart-monai . \
        && rm -rf _tmp \
        && cd quickstart-monai

This will create a new directory called quickstart-monai with the following structure:

quickstart-monai
├── monaiexample
│   ├── __init__.py
│   ├── client_app.py   # Defines your ClientApp
│   ├── server_app.py   # Defines your ServerApp
│   └── task.py         # Defines your model, training and data loading
├── pyproject.toml      # Project metadata like dependencies and configs
└── README.md

Install dependencies and project

Install the dependencies defined in pyproject.toml as well as the monaiexample package.

pip install -e .

Run the project

You can run your Flower project in both simulation and deployment mode without making changes to the code. If you are starting with Flower, we recommend you using the simulation mode as it requires fewer components to be launched manually. By default, flwr run will make use of the Simulation Engine.

Run with the Simulation Engine

[!TIP] This example runs faster when the ClientApps have access to a GPU. If your system has one, you can make use of it by configuring the backend.client-resources component in pyproject.toml. If you want to try running the example with GPU right away, use the local-simulation-gpu federation as shown below. Check the Simulation Engine documentation to learn more about Flower simulations and how to optimize them.

# Run with the default federation (CPU only)
flwr run .

Run the project in the local-simulation-gpu federation that gives CPU and GPU resources to each ClientApp. By default, at most 4xClientApp will run in parallel in the available GPU.

# Run with the `local-simulation-gpu` federation
flwr run . local-simulation-gpu

You can also override some of the settings for your ClientApp and ServerApp defined in pyproject.toml. For example:

flwr run . --run-config "num-server-rounds=5 batch-size=32"

Run with the Deployment Engine

Follow this how-to guide to run the same app in this example but with Flower’s Deployment Engine. After that, you might be intersted in setting up secure TLS-enabled communications and SuperNode authentication in your federation.

If you are already familiar with how the Deployment Engine works, you may want to learn how to run it using Docker. Check out the Flower with Docker documentation.