Image-Text Retrieval on CBVS-20K (test)
50.3R@1UniCLIP
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| UniCLIPMode=Fine-tuning, Backbone=ViT-B/162024.01 | 50.3 | 75.4 | 82 | 0.692 | 3.069 | 0.784 | 0.846 | 0.893 | 0.779 | |
| QA-CLIPMode=Fine-tuning, Backbone=ViT-B/162024.01 | 47.3 | 71.1 | 78.3 | 0.656 | 2.907 | 0.778 | 0.841 | 0.89 | 0.771 | |
| CN-CLIPMode=Fine-tuning, Backbone=ViT-B/162024.01 | 47.1 | 72.1 | 79.6 | 0.663 | 2.914 | 0.767 | 0.838 | 0.888 | 0.767 | |
| ALBEF-CLIPMode=Fine-tuning, Backbone=ViT-B/162024.01 | 46.8 | 73.1 | 79.4 | 0.664 | 2.906 | 0.771 | 0.839 | 0.889 | 0.769 | |
| CN-CLIPMode=Zero-shot, Backbone=ViT-L/142024.01 | 43.4 | 68.5 | 75.6 | 0.625 | 2.812 | 0.773 | 0.84 | 0.889 | 0.775 | |
| R2D2-250MMode=Fine-tuning, Backbone=ViT-L/142024.01 | 41.8 | 60.5 | 65 | 0.558 | 2.934 | 0.78 | 0.844 | 0.891 | 0.774 | |
| Ernie-ViL2.0Mode=Zero-shot, Backbone=ViT-B2024.01 | 41.3 | 66 | 74.2 | 0.605 | 2.759 | 0.764 | 0.835 | 0.886 | 0.768 | |
| QA-CLIPMode=Zero-shot, Backbone=ViT-B/162024.01 | 40 | 65.2 | 72.4 | 0.592 | 2.804 | 0.774 | 0.838 | 0.888 | 0.77 | |
| CN-CLIPMode=Zero-shot, Backbone=ViT-B/162024.01 | 38.4 | 62.8 | 70.4 | 0.572 | 2.718 | 0.768 | 0.835 | 0.885 | 0.764 | |
| R2D2-250MMode=Zero-shot, Backbone=ViT-L/142024.01 | 35.6 | 51.2 | 53.5 | 0.468 | 2.829 | 0.789 | 0.842 | 0.891 | 0.775 | |
| WuKongMode=Zero-shot, Backbone=ViT-L/142024.01 | 31.1 | 50.3 | 58.3 | 0.466 | 2.229 | 0.739 | 0.811 | 0.872 | 0.738 | |
| Taiyi-CLIPMode=Zero-shot, Backbone=ViT-L2024.01 | 26.9 | 49.2 | 57.7 | 0.446 | 2.278 | 0.726 | 0.805 | 0.866 | 0.733 | |
| R2D2-23MMode=Zero-shot, Backbone=ViT-L/142024.01 | 25.8 | 40.7 | 43.6 | 0.367 | 2.312 | 0.733 | 0.81 | 0.868 | 0.738 | |
| Taiyi-CLIPMode=Zero-shot, Backbone=ViT-B2024.01 | 25.1 | 44.5 | 52.5 | 0.407 | 2.142 | 0.718 | 0.8 | 0.861 | 0.727 | |
| WuKongMode=Zero-shot, Backbone=ViT-B/322024.01 | 19.7 | 35.6 | 43.9 | 0.331 | 2 | 0.702 | 0.791 | 0.858 | 0.712 | |
| AltCLIPMode=Zero-shot, Backbone=ViT-L2024.01 | 16.2 | 28.4 | 33.6 | 0.261 | 1.869 | 0.669 | 0.771 | 0.842 | 0.701 |