FedAvP: Augmenting Local Data via Shared Policy in Federated Learning

This repository contains the official PyTorch implementation for the paper:

• Minui Hong, Junhyeog Yun, Insu Jeon, and Gunhee Kim. "FedAvP: Augmenting Local Data via Shared Policy in Federated Learning." NeurIPS 2024. [paper]

Note: This repository will be further updated soon.

Requirements

• Python 3.7.12
• PyTorch 1.13.1

Example Usage (CIFAR-100 / α=0.1)

The dataset will be automatically downloaded when you run the code.

To train and evaluate FedAvP (FastUpdate):

python fedavp_FastUpdate.py

To train and evaluate FedAvP:

python fedavp.py

Acknowledgments

This code is adapted from the following implementations:

• https://github.com/AIoT-MLSys-Lab/DeepAA
• https://github.com/ildoonet/pytorch-randaugment/blob/master/RandAugment/augmentations.py
• https://github.com/insujeon/MetaVD