Multi-class Classification on FACET
56.37AccuracyPrompt-D
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Prompt-DModel=XVLM, Evaluation protocol=Zero-shot2025.11 | 56.37 | 10.35 | 11.7 | |
| DeARModel=XVLM, Evaluation protocol=Zero-shot2025.11 | 56.3 | 11.26 | 3.9 | |
| BaselineModel=XVLM, Evaluation protocol=Zero-shot2025.11 | 55.74 | 11.72 | — | |
| CLIP-clipModel=XVLM, Evaluation protocol=Zero-shot2025.11 | 54.52 | 9.98 | 14.8 | |
| SPD (Ours)Model=XVLM, Evaluation protocol=Zero-shot2025.11 | 54.32 | 9.85 | 16 | |
| SFIDModel=XVLM, Evaluation protocol=Zero-shot2025.11 | 53.69 | 9.91 | 15.4 | |
| Prompt-DModel=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 52.58 | 10.37 | 6.4 | |
| BaselineModel=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 52.17 | 11.6 | — | |
| SFIDModel=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 52.14 | 10.15 | 12.5 | |
| DeARModel=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 52.08 | 10.04 | 9.4 | |
| Prompt-DModel=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 51.96 | 10.56 | 9 | |
| BaselineModel=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 51.87 | 11.08 | — | |
| CLIP-clipModel=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 51.56 | 10.8 | 6.9 | |
| SPD (Ours)Model=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 51.44 | 9.55 | 13.8 | |
| SPD (Ours)Model=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 51.29 | 9.94 | 14.3 | |
| SFIDModel=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 50.93 | 9.63 | 13.1 | |
| CLIP-clipModel=CLIP (ResNet50), Evaluation protocol=Zero-shot2025.11 | 50.73 | 10.09 | 8.9 | |
| DeARModel=CLIP (ViT-B/32), Evaluation protocol=Zero-shot2025.11 | 50.09 | 10.37 | 10.6 |