/federated_learning

Implementation of federated learning in python

Primary LanguagePython

Federated Learning Framework

This is an easy to use framework for federated learning along with a react ui for the same.

User Interface

To get started follow the steps below:

  1. Install pytorch, torchvision for cpu or gpu as required

  2. Run pip install -r requirements.txt to install the common requirements

  3. To install necessary files for react UI:

    • Install NodeJs.
    • cd react-app
    • npm install

  4. Use command npm start to start the UI. Head to localhost:3000 to view the UI

  5. Change server_config.yaml to suit your needs

  6. Have a copy of the client folder at each client and change client_config.yaml for each separate client,

  7. Make sure to mention the dataset paths,labels, etc

  8. Start the server by running python run_new_server.py

  9. To start a client, cd to the client directory and run python run_new_client.py to start a new client

If a client crashes, run 'python main_client.py' to resume the client and continue from the most recent fl round.