ss3b3/FedAWARE

Code for paper - Tackling Hybrid Heterogeneity on Federated Optimization via Gradient Diversity Maximization

FedAWARE

Code for paper "Tackling Hybrid Heterogeneity on Federated Optimization via Gradient Diversity Maximization".

Dependencies

pip install -r requirement.txt

Run

python fedaware.py -num_clients 100 \
                    -com_round 500 \
                    -sample_ratio 0.1 \
                    -batch_size 64 \
                    -epochs 3 \
                    -lr 0.01 \
                    -glr 1 \
                    -dseed 37 \
                    -seed 1998 \
                    -partition [pathological/dirichlet] \
                    -dataset [mnist\fmnist\cifar10] \
                    -alpha 0.5 \
                    -startup 1 \
                    -agnostic [0\1] \
                    -preprocess 1

Citation

Please cite our paper if you found the code useful.

@article{zeng2023tackling,
  title={Tackling Hybrid Heterogeneity on Federated Optimization via Gradient Diversity Maximization},
  author={Zeng, Dun and Xu, Zenglin and Pan, Yu and Wang, Qifan and Tang, Xiaoying},
  journal={arXiv preprint arXiv:2310.02702},
  year={2023}
}