APPFL: Argonne Privacy-Preserving Federated Learning

_images/logo_small.png

APPFL is a privacy-preserving federated learning framework that allows users to implement a federated learning environment with

  • user-defined neural network models (based on torch.nn.Module)

  • customized algorithms

  • privacy techniques.

The framework is designed to run on a single machine (a laptop or a cluster) as well as multiple heterogeneous machines.

Table of Contents

Getting Started

Users Guide

Developers Guide

Acknowledgement

This material is based upon work supported by the U.S. Department of Energy, Office of Science, under contract number DE-AC02-06CH11357.