Composed Image Retrieval on GeneCIS (Full R@K Specialized Metrics)
22.9Focus Attribute R@1SEIZE
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SEIZEArch=ViT-G/14, Train=No2026.05 | 22.9 | 36.2 | 47.3 | 18.6 | 31.4 | 42.7 | 18.2 | 28.8 | 37.6 | 19.6 | 33 | 43.5 | 19.8 | |
| SEIZEArch=ViT-G/14, Train=false2026.05 | 22.9 | 36.2 | 47.3 | 18.6 | 31.4 | 42.7 | 18.2 | 28.8 | 37.6 | 19.6 | 33 | 43.5 | 19.8 | |
| OSrCIRArch=ViT-G/14, Train=No2026.05 | 22.7 | 36.4 | 47 | 17.9 | 30.8 | 42 | 16.9 | 28.4 | 36.7 | 21 | 33.4 | 44.2 | 19.6 | |
| OSrCIRArch=ViT-G/14, Train=false2026.05 | 22.7 | 36.4 | 47 | 17.9 | 30.8 | 42 | 16.9 | 28.4 | 36.7 | 21 | 33.4 | 44.2 | 19.6 | |
| STiTchArch=ViT-G/14, Train=No2026.05 | 21.9 | 36.4 | 47.9 | 19.6 | 31.9 | 42.8 | 20.2 | 30.3 | 39.6 | 19.7 | 33.2 | 43.4 | 20.4 | |
| STiTchArch=ViT-G/14, Train=false2026.05 | 21.9 | 36.4 | 47.9 | 19.6 | 31.9 | 42.8 | 20.2 | 30.3 | 39.6 | 19.7 | 33.2 | 43.4 | 20.4 | |
| STiTchArch=ViT-B/32, Train=false2026.05 | 21.1 | 35 | 45.5 | 17.9 | 29.9 | 40.4 | 16.4 | 28.5 | 38.9 | 18.3 | 30.1 | 39.5 | 18.4 | |
| OSrCIRArch=ViT-L/14, Train=No2026.05 | 20.9 | 33.1 | 44.5 | 17.2 | 28.5 | 37.9 | 15 | 23.6 | 34.2 | 18.4 | 30.6 | 38.3 | 17.9 | |
| CIReVLArch=ViT-G/14, Train=No2026.05 | 20.9 | 34.4 | 44.9 | 16.5 | 29 | 39.8 | 15.1 | 25.6 | 33.4 | 18.5 | 31.6 | 41.4 | 17.8 | |
| OSrCIRArch=ViT-L/14, Train=false2026.05 | 20.9 | 33.1 | 44.5 | 17.2 | 28.5 | 37.9 | 15 | 23.6 | 34.2 | 18.4 | 30.6 | 38.3 | 17.9 | |
| SEIZEArch=ViT-L/14, Train=No2026.05 | 20.5 | 33.4 | 45 | 17.6 | 28.9 | 38.5 | 15.4 | 25.6 | 36.2 | 18.7 | 30.9 | 39.8 | 18.1 | |
| SEIZEArch=ViT-L/14, Train=false2026.05 | 20.5 | 33.4 | 45 | 17.6 | 28.9 | 38.5 | 15.4 | 25.6 | 36.2 | 18.7 | 30.9 | 39.8 | 18.1 | |
| CIReVLArch=ViT-G/14, Train=false2026.05 | 20.5 | 34 | 44.5 | 16.1 | 28.6 | 39.4 | 14.7 | 25.2 | 33 | 18.1 | 31.2 | 41 | 17.4 | |
| STiTchArch=ViT-L/14, Train=No2026.05 | 20.3 | 34.6 | 46.4 | 18.3 | 29.8 | 41.6 | 16.8 | 28.5 | 38.4 | 18.8 | 31 | 40.3 | 18.6 | |
| STiTchArch=ViT-L/14, Train=false2026.05 | 20.3 | 34.6 | 46.4 | 18.3 | 29.8 | 41.6 | 16.8 | 28.5 | 38.4 | 18.8 | 31 | 40.3 | 18.6 | |
| CIReVArch=ViT-L/14, Train=No2026.05 | 19.5 | 31.8 | 42 | 14.4 | 26 | 35.2 | 12.3 | 21.8 | 30.5 | 17.2 | 28.9 | 37.6 | 15.9 | |
| CIReVLArch=ViT-L/14, Train=false2026.05 | 19.5 | 31.8 | 42 | 14.4 | 26 | 35.2 | 12.3 | 21.8 | 30.5 | 17.2 | 28.9 | 37.6 | 15.9 | |
| OSrCIRArch=ViT-B/32, Train=false2026.05 | 19.4 | 32.7 | 42.8 | 16.4 | 27.7 | 38.1 | 15.7 | 25.7 | 35.8 | 18.2 | 30.1 | 39.4 | 17.4 | |
| LinCIRArch=ViT-G/14, Train=true2026.05 | 19.1 | 33 | 42.3 | 17.6 | 30.2 | 38.1 | 10.1 | 19.1 | 28.1 | 7.9 | 16.3 | 25.7 | 13.7 | |
| SEARLEArch=ViT-B/32, Train=true2026.05 | 18.9 | 30.6 | 41.2 | 13 | 23.8 | 33.7 | 12.2 | 23 | 33.3 | 13.6 | 23.8 | 33.3 | 14.4 | |
| CIReVLArch=ViT-B/32, Train=false2026.05 | 17.9 | 29.4 | 40.4 | 14.8 | 25.8 | 35.8 | 14.6 | 24.3 | 33.3 | 16.1 | 27.8 | 37.6 | 15.9 | |
| Context-I2WArch=ViT-L/14, Train=Yes2026.05 | 17.2 | 30.5 | 41.7 | 16.4 | 28.3 | 37.1 | 8.7 | 17.9 | 26.9 | 7.7 | 16 | 25.4 | 12.7 | |
| Context-I2WArch=ViT-L/14, Train=true2026.05 | 17.2 | 30.5 | 41.7 | 16.4 | 28.3 | 37.1 | 8.7 | 17.9 | 26.9 | 7.7 | 16 | 25.4 | 12.7 | |
| SEARLEArch=ViT-L/14, Train=Yes2026.05 | 17.1 | 29.6 | 40.7 | 16.3 | 25.2 | 34.2 | 12 | 22.2 | 30.9 | 12 | 24.1 | 33.9 | 14.4 | |
| SEARLEArch=ViT-L/14, Train=true2026.05 | 17.1 | 29.6 | 40.7 | 16.3 | 25.2 | 34.2 | 12 | 22.2 | 30.9 | 12 | 24.1 | 33.9 | 14.4 | |
| LinCIRArch=ViT-L/14, Train=Yes2026.05 | 16.9 | 30 | 41.5 | 16.2 | 28 | 36.8 | 8.3 | 17.4 | 26.2 | 7.4 | 15.7 | 25 | 12.2 | |
| LinCIRArch=ViT-L/14, Train=true2026.05 | 16.9 | 30 | 41.5 | 16.2 | 28 | 36.8 | 8.3 | 17.4 | 26.2 | 7.4 | 15.7 | 25 | 12.2 |