Image-to-Text Retrieval on UCM-Caption (test)
20.48Recall@1LRSCLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LRSCLIPBackbone=ViT-B/162026.03 | 20.48 | 52.86 | 82.38 | |
| GeoAlignCLIPBackbone=ViT-L/142026.03 | 20.48 | 52.86 | 77.62 | |
| GeoRSCLIPBackbone=ViT-L/14-3362026.03 | 19.05 | 49.05 | 70.95 | |
| LRSCLIPBackbone=ViT-L/142026.03 | 19.05 | 52.86 | 77.14 | |
| GeoRSCLIPBackbone=ViT-L/142026.03 | 18.57 | 46.19 | 71.9 | |
| FGCLIPBackbone=ViT-L/14-3362026.03 | 17.62 | 48.57 | 69.05 | |
| GeoRSCLIPBackbone=ViT-B/322026.03 | 16.67 | 44.76 | 72.38 | |
| GeoAlignCLIPBackbone=ViT-B/162026.03 | 15.71 | 49.52 | 74.29 | |
| SkyCLIPBackbone=ViT-B/322026.03 | 15.24 | 40.95 | 63.81 | |
| SkyCLIPBackbone=ViT-L/14-pct302026.03 | 15.24 | 44.76 | 66.67 | |
| SkyCLIPBackbone=ViT-L/14-pct502026.03 | 14.76 | 44.76 | 68.1 | |
| LongCLIPBackbone=ViT-L/142026.03 | 13.33 | 41.9 | 67.14 | |
| LongCLIPBackbone=ViT-B/162026.03 | 12.86 | 41.9 | 65.24 | |
| CLIPBackbone=ViT-L/142026.03 | 12.38 | 41.43 | 70.95 | |
| FGCLIPBackbone=ViT-B/162026.03 | 11.9 | 38.1 | 65.71 | |
| CLIPBackbone=ViT-B/162026.03 | 8.57 | 31.9 | 55.71 |