Text-to-Image Retrieval on UCM-Caption (test)
17.71Recall@1GeoAlignCLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GeoAlignCLIPBackbone=ViT-L/142026.03 | 17.71 | 55.33 | 86.1 | |
| GeoRSCLIPBackbone=ViT-L/142026.03 | 16.1 | 54.48 | 86.19 | |
| LRSCLIPBackbone=ViT-L/142026.03 | 15.71 | 55.81 | 85.24 | |
| FGCLIPBackbone=ViT-L/14-3362026.03 | 15.43 | 55.62 | 84 | |
| GeoRSCLIPBackbone=ViT-L/14-3362026.03 | 15.43 | 54.57 | 84.67 | |
| GeoRSCLIPBackbone=ViT-B/322026.03 | 14.76 | 51.05 | 81.81 | |
| LRSCLIPBackbone=ViT-B/162026.03 | 14.29 | 53.24 | 85.62 | |
| GeoAlignCLIPBackbone=ViT-B/162026.03 | 13.81 | 50.57 | 79.43 | |
| LongCLIPBackbone=ViT-L/142026.03 | 12.95 | 49.62 | 81.05 | |
| FGCLIPBackbone=ViT-B/162026.03 | 12.57 | 47.24 | 75.9 | |
| SkyCLIPBackbone=ViT-L/14-pct302026.03 | 12.48 | 47.24 | 76.76 | |
| SkyCLIPBackbone=ViT-L/14-pct502026.03 | 12.19 | 46.86 | 76.29 | |
| SkyCLIPBackbone=ViT-B/322026.03 | 11.9 | 46.86 | 77.33 | |
| LongCLIPBackbone=ViT-B/162026.03 | 11.81 | 45.81 | 76.19 | |
| CLIPBackbone=ViT-L/142026.03 | 10.57 | 43.81 | 71.24 | |
| CLIPBackbone=ViT-B/162026.03 | 9.52 | 36 | 64.1 |