Multi-class Classification on FACET (test)
56.37AccuracyPrompt-Debias
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Prompt-DebiasBackbone=XVLM, Zero-shot=true2024.10 | 56.37 | 10.35 | |
| DeARBackbone=XVLM, Zero-shot=true2024.10 | 56.3 | 11.26 | |
| BaselineBackbone=XVLM, Zero-shot=true2024.10 | 55.74 | 11.72 | |
| CLIP-clipBackbone=XVLM, Zero-shot=true2024.10 | 54.52 | 9.98 | |
| SFIDBackbone=XVLM, Zero-shot=true2024.10 | 53.69 | 9.91 | |
| Prompt-DebiasBackbone=ResNet-50, Zero-shot=true2024.10 | 52.58 | 10.37 | |
| BaselineBackbone=ViT-B/32, Zero-shot=true2024.10 | 52.17 | 11.6 | |
| SFIDBackbone=ViT-B/32, Zero-shot=true2024.10 | 52.14 | 10.15 | |
| DeARBackbone=ResNet-50, Zero-shot=true2024.10 | 52.08 | 10.04 | |
| Prompt-DebiasBackbone=ViT-B/32, Zero-shot=true2024.10 | 51.96 | 10.56 | |
| BaselineBackbone=ResNet-50, Zero-shot=true2024.10 | 51.87 | 11.08 | |
| CLIP-clipBackbone=ViT-B/32, Zero-shot=true2024.10 | 51.56 | 10.8 | |
| SFIDBackbone=ResNet-50, Zero-shot=true2024.10 | 50.93 | 9.63 | |
| CLIP-clipBackbone=ResNet-50, Zero-shot=true2024.10 | 50.73 | 10.09 | |
| DeARBackbone=ViT-B/32, Zero-shot=true2024.10 | 50.09 | 10.37 |