Skip to content

An easy implementation of a federated learning scenario on CNN

License

Notifications You must be signed in to change notification settings

Ls-Dai/Pytorch_FL_CNN

Repository files navigation

Pytorch_FL_CNN

An easy implementation of a federated learning scenario on CNN

There are some discription about my code

clear_logs.py

This will remove all the data of client folders, server folders and a log file in pwd. Make all records clear.

client.py

This is the module that contains a Client() class, which will be used as client module in FL.

configs.py

A module with a class that contains hyperparameters in both model part and traing part.

init.py

Module that would initialize a FL (assign servers and clients, and load dataset).

pip_update_all.py

This will update all packages of pip.

server.py

The module that contains a Server() class, which will be used as server module in FL.

test.py

Deprecated. It was used to test a FL framework before.

train.py

Load configurations, clients, servers. Initialize a FL framework and train it.

train_process_vis.py

This can read log files and visualize the training process after your training is completed.

utils.py

Deprecated. All methods have been moved to client and server modules.

About

An easy implementation of a federated learning scenario on CNN

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages