Federated Learning Classification on CIFAR-100 non-iid delta=0.3
55.27AccuracyFedDyn+MAN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FedDyn+MANRegularizer=MAN2025.12 | 55.27 | 145 | |
| FedSpeed+MANRegularizer=MAN2025.12 | 55.23 | 163 | |
| FedDC+MANRegularizer=MAN2025.12 | 55.21 | 144 | |
| FedDCRegularizer=None2025.12 | 52.02 | 294 | |
| FedAvg+MANRegularizer=MAN2025.12 | 52 | 206 | |
| FedSAM+MANRegularizer=MAN2025.12 | 51.59 | 326 | |
| FedASAM+MANRegularizer=MAN2025.12 | 51.23 | 313 | |
| FedSpeedRegularizer=None2025.12 | 50.95 | 392 | |
| FedDynRegularizer=None2025.12 | 49.29 | 500 | |
| FedASAMRegularizer=None2025.12 | 46 | 500 | |
| FedSAMRegularizer=None2025.12 | 43.44 | 500 | |
| FedAvgRegularizer=None2025.12 | 40.9 | 500 |