Image Classification on FMNIST (label shift, test)
72.66Top-1 AccuracyLSS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LSSBackbone=ResNet-18, Pre-trained=ImageNet, Memory=4x2024.10 | 72.66 | 92.45 | |
| DiWABackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 63.21 | 91.88 | |
| SoupsBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 60.11 | 91.56 | |
| SWADBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 58.66 | 91.22 | |
| SWABackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 55.82 | 91.03 | |
| FedRepBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 36.2 | 91.07 | |
| FedBABUBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 36.18 | 91.31 | |
| MOONBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 36.01 | 91.28 | |
| FedAvgBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 35.54 | 90.04 | |
| FedBNBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 34.2 | 89.87 | |
| FedFomoBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 33.94 | 88.41 | |
| FedProxBackbone=ResNet-18, Pre-trained=ImageNet2024.10 | 33.48 | 89.28 |