This is the official implementation of the CVPR 2025 paper "FedAWA: Adaptive Optimization of Aggregation Weights in Federated Learning Using Client Vectors".
Our code is based on Python version 3.7 and PyTorch version 1.13.1. You can run FedAWA with the following command:
python main.py --dataset cifar100 --local_model ResNet20 --server_method fedawa --client_method local_train #FedAWA on CIFAR-100 dataset with ResNet20 model
Please cite our paper if you find this repo useful in your work:
@misc{shi2025fedawaadaptiveoptimizationaggregation,
title={FedAWA: Adaptive Optimization of Aggregation Weights in Federated Learning Using Client Vectors},
author={Changlong Shi and He Zhao and Bingjie Zhang and Mingyuan Zhou and Dandan Guo and Yi Chang},
year={2025},
eprint={2503.15842},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2503.15842},
}
We would like to thank the authors for releasing the public repository: ICML-2023-FedLAW
Please feel free to contact via email ([email protected]) if you have any questions.