Image Classification on Caltech101 (Zero-shot)
93.6Clean AccuracyGRACE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GRACEBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 93.6 | 72.8 | |
| TPGMBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 93.5 | 0 | |
| Vanilla FTBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 93.2 | 0 | |
| FLYPBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 93.1 | 0 | |
| SPDBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 93 | 0 | |
| WISE-FTBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 92.8 | 0 | |
| CLIPBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 91.48 | 0.41 | |
| FAREBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 88.2 | 42.5 | |
| PMG-AFTBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 86.5 | 71.2 | |
| LAATBackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 85.8 | 68.4 | |
| TeCoABackbone=ViT-B/16, Training=ImageNet, Evaluation=Zero-shot2026.03 | 84.63 | 69.07 |