Skip to content

jopasserat/federated-learning-tutorial

Repository files navigation

FL simulation with medical imaging classification task

This code splits the Pathology MedMNIST dataset into pool_size partitions (user defined) and does a few rounds of training.

Requirements

  • Flower 0.19.0
  • A recent version of PyTorch. This example has been tested with Pytorch 1.11.0
  • A recent version of Ray. This example has been tested with Ray 1.11.1.

Install

Create a new Conda environment with Python 3.9, the following commands will isntall all the dependencies needed:

conda create --name my_project_env --file conda-linux-64.lock
poetry install

Updating the environment

# Re-generate Conda lock file(s) based on environment.yml
conda-lock -k explicit --conda mamba
# Update Conda packages based on re-generated lock file
mamba update --file conda-linux-64.lock
# Update Poetry packages and re-generate poetry.lock
poetry update

How to run

This example:

  1. Downloads Pathology MedMNIST
  2. Partitions the dataset into N splits, where N is the total number of clients. We refere to this as pool_size. The partition can be IID or non-IID
  3. Starts a Ray-based simulation where a % of clients are sample each round. This example uses N=3, so 3 clients will be sampled each round.
  4. After the M rounds end, the global model is evaluated on the entire testset. Also, the global model is evaluated on the valset partition residing in each client. This is useful to get a sense on how well the global model can generalise to each client's data.

The command below will assign each client 1 CPU threads. If your system does not have 1xN(=3) = 3 threads to run all 3 clients in parallel, they will be queued but eventually run. The server will wait until all N clients have completed their local training stage before aggregating the results. After that, a new round will begin.

$ python main.py --num_client_cpus 2 # note that `num_client_cpus` should be <= the number of threads in your system.

References