Image Classification on DomainNet alpha=0.1 (test)
70Accuracy (Clipart)FedCGS + SLOT-Align
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| FedCGS + SLOT-AlignBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot, Alignment strength (tau)=0.42026.06 | 70 | 40.59 | 63.02 | 28.54 | 75.42 | 62.21 | 56.63 | 16.59 | |
| FedCGSBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot2026.06 | 66.67 | 31.74 | 58.16 | 20.66 | 73.08 | 58.11 | 51.4 | 18.82 | |
| FedPFT + SLOT-AlignBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot, Alignment strength (tau)=0.42026.06 | 65.54 | 34.83 | 57.45 | 26.73 | 68.37 | 56.48 | 51.57 | 15.46 | |
| FedAvgBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=Multi-round2026.06 | 64.05 | 30.37 | 53.81 | 25.58 | 64.28 | 53.18 | 48.1 | 15.26 | |
| FedPFTBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot2026.06 | 62.85 | 29.13 | 54.71 | 23.54 | 65.86 | 52.13 | 48.04 | 16.1 | |
| O-FedAvgBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot2026.06 | 56.93 | 24.55 | 43.34 | 17.56 | 53.09 | 40.69 | 37.69 | 12.53 | |
| O-FedAvg + SLOT-AlignBackbone=ViT-B/32, Pre-training=CLIP, Federated Learning Rounds=One-shot, Alignment strength (tau)=0.42026.06 | 49.9 | 27.73 | 45.16 | 20.03 | 54.56 | 43.91 | 40.26 | 12.18 |