Text-to-Image Retrieval on AIC-ICC (test)
57.3R@1R2D2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| R2D2Backbone=ViT-L2022.05 | 57.3 | 78.1 | 83 | — | |
| WukongBackbone=ViT-L2022.05 | 48.6 | 72.5 | 80.2 | — | |
| R2D2Backbone=ViT-B2022.05 | 47.6 | 72.8 | 80.2 | — | |
| CLIPBackbone=ViT-L2022.05 | 46.2 | 70.7 | 78.6 | — | |
| FILIPBackbone=ViT-L2022.05 | 44.9 | 69 | 77.5 | — | |
| CLIPBackbone=ViT-B2022.05 | 38.1 | 63.7 | 73.3 | — | |
| WukongBackbone=ViT-B2022.05 | 36.7 | 36.7 | 71.7 | — | |
| BriVL2022.05 | 34.1 | 58.9 | 69.1 | — | |
| FILIPBackbone=ViT-B2022.05 | 32.9 | 58.4 | 68.8 | — |