This is official PyTorch implementation of the paper "Does one-shot give the best shot? Mitigating Model Inconsistency in One-shot Federated Learning" (accepted by ICML 2025).
- python 3.8 (Anaconda)
- Pytorch 1.10.1
- torchvision 0.11.2
- CUDA 11.4
- CIFAR-10
- CIFAR-100
- SVHN
- Tiny-ImageNet
Commands can be found in run.sh
All configurations are in the folder of `configs'
@inproceedings{fafi,
author = {Hui, Zeng and Wenke, Huang and Tongqing, Zhou and Xinyi, Wu and Guancheng, Wan and Yingwen, Chen and Zhiping, Cai},
title = {Does one-shot give the best shot? Mitigating Model Inconsistency in One-shot Federated Learning},
year = {2025},
booktitle = {Proceedings of International Conference on Machine Learning (ICML)},
numpages = {9},
}