Image Classification on Tiny-ImageNet (Directional Accuracy Metrics)
54.06Accuracy (Dir-0.6)FedACT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FedACTBackbone=ResNet-18, Communication Rounds=3002026.07 | 54.06 | 53.62 | 51.09 | |
| FedLADABackbone=ResNet-18, Communication Rounds=3002026.07 | 54.04 | 53.01 | 50.5 | |
| SCAFFOLDBackbone=ResNet-18, Communication Rounds=3002026.07 | 53.78 | 52.41 | 48.22 | |
| FedAdamWBackbone=ResNet-18, Communication Rounds=3002026.07 | 53.66 | 52.71 | 50.65 | |
| FedAvgBackbone=ResNet-18, Communication Rounds=3002026.07 | 53.59 | 51.79 | 48.38 | |
| FedProxBackbone=ResNet-18, Communication Rounds=3002026.07 | 53.07 | 51.46 | 47.69 | |
| FedAdamBackbone=ResNet-18, Communication Rounds=3002026.07 | 50.87 | 49.61 | 45.36 | |
| LocalAdamBackbone=ResNet-18, Communication Rounds=3002026.07 | 50.82 | 48.99 | 45.13 | |
| LocalAdamWBackbone=ResNet-18, Communication Rounds=3002026.07 | 50.62 | 49.46 | 44.95 |