https://github.com/vivekkhimani/federated_learning_pysyft

A simple federated learning implementation on MNIST dataset using PySyft. Main goal of the project was to get used to the PySyft federated learning functionality instead of using traditional PyTorch features.
https://github.com/vivekkhimani/federated_learning_pysyft

classification-model federated-learning mnist pysyft pytorch workers

Last synced: about 2 months ago
JSON representation

A simple federated learning implementation on MNIST dataset using PySyft. Main goal of the project was to get used to the PySyft federated learning functionality instead of using traditional PyTorch features.

• Host: GitHub
• URL: https://github.com/vivekkhimani/federated_learning_pysyft
• Owner: vivekkhimani
• License: mit
• Created: 2019-11-12T22:33:27.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2024-04-19T07:05:40.000Z (about 1 year ago)
• Last Synced: 2025-03-24T17:52:20.578Z (2 months ago)
• Topics: classification-model, federated-learning, mnist, pysyft, pytorch, workers
• Language: Python
• Homepage:
• Size: 8.79 KB
• Stars: 13
• Watchers: 1
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Implementing Federated Learning using PySyft

### Basics:

- Dataset - MNIST

- Number of Workers - 32

- Classification Model - CNN (see the details in models directory)

- Tools Used - PySyft, PyTorch

### Instructions:

- Prerequisite: python3, pip3, pysyft, pytorch

- RUN: "main_fed.py"

- To edit the basic characteristics of the model, check "/utils/Arguments.py". No CLI has been provided for now. 

- To edit the classification model, check "/models/CNN.py"

### Future Work:

- Add a CLI to make the process of editing the arguments easier. 

- Facilitate training by selecting a subset of workers instead of using all the workers.