Image Classification on CIFAR100 (ACC, CEA)
72.1AccuracyPS-CBM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PS-CBMBackbone=CLIP RN502025.11 | 72.1 | 53.6 | |
| Linear ProbeBackbone=CLIP RN502025.11 | 69.8 | — | |
| LF-CBMBackbone=CLIP RN502025.11 | 68.8 | 49.9 | |
| DN-CBMBackbone=CLIP RN502025.11 | 68.6 | 46.7 | |
| VLG-CBMBackbone=CLIP RN502025.11 | 68.3 | 50.8 | |
| V2C-CBMBackbone=CLIP RN502025.11 | 68.3 | 47.2 | |
| LaBoBackbone=CLIP RN502025.11 | 68.1 | 47.1 | |
| LM4CVBackbone=CLIP RN502025.11 | 65.9 | 49.3 | |
| Res-CBMBackbone=CLIP RN502025.11 | 65.7 | 49.6 | |
| DCBMBackbone=CLIP RN502025.11 | 62.3 | 44.3 |