Text-to-Image Retrieval on LIT-CN
86.3Recall@1FG-CLIP 2
Evaluation results
| Method | Links | |
|---|---|---|
| FG-CLIP 2Backbone=ViT-So/162025.10 | 86.3 | |
| FG-CLIP 2Backbone=ViT-L/162025.10 | 85.9 | |
| FG-CLIP 2Backbone=ViT-B/162025.10 | 81.1 | |
| Meta CLIP 2Backbone=ViT-H/142025.10 | 67.6 | |
| Chinese-CLIPBackbone=ViT-L/142025.10 | 38.9 | |
| Chinese-CLIPBackbone=ViT-B/162025.10 | 35.6 | |
| R2D2Backbone=ViT-L/142025.10 | 33.3 | |
| R2D2Backbone=ViT-B/162025.10 | 27.4 | |
| SigLIP 2Backbone=ViT-L/162025.10 | 10.9 | |
| SigLIP 2Backbone=ViT-So/162025.10 | 10.8 | |
| SigLIP 2Backbone=ViT-B/162025.10 | 2.6 |