Text-to-Image Retrieval on DCI-CN
65.1Recall@1FG-CLIP 2
Evaluation Results
| Method | Links | |
|---|---|---|
| FG-CLIP 2Backbone=ViT-So/162025.10 | 65.1 | |
| FG-CLIP 2Backbone=ViT-L/162025.10 | 62.2 | |
| FG-CLIP 2Backbone=ViT-B/162025.10 | 55.7 | |
| Meta CLIP 2Backbone=ViT-H/142025.10 | 52.1 | |
| R2D2Backbone=ViT-L/142025.10 | 34.2 | |
| Chinese-CLIPBackbone=ViT-L/142025.10 | 32.7 | |
| Chinese-CLIPBackbone=ViT-B/162025.10 | 27.9 | |
| R2D2Backbone=ViT-B/162025.10 | 27.3 | |
| SigLIP 2Backbone=ViT-L/162025.10 | 14.4 | |
| SigLIP 2Backbone=ViT-So/162025.10 | 12 | |
| SigLIP 2Backbone=ViT-B/162025.10 | 4 |