Composed Image Retrieval on Fashion-IQ Average
59.09Recall@10IMAGINE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| IMAGINEModel Category=CVR Models2026.06 | 59.09 | 80.11 | |
| LIMN+Model Category=CIR Models2026.06 | 57.43 | 78.6 | |
| ENCODERModel Category=CIR Models2026.06 | 56.13 | 77.59 | |
| LIMNModel Category=CIR Models2026.06 | 55.91 | 77.82 | |
| SPRCModel Category=CIR Models2026.06 | 54.72 | 74.97 | |
| PAIRModel Category=CIR Models2026.06 | 52.76 | 74.51 | |
| QuReModel Category=CIR Models2026.06 | 52.6 | 73.48 | |
| MEDIANModel Category=CIR Models2026.06 | 52.39 | 74.3 | |
| TG-CIRModel Category=CIR Models2026.06 | 51.32 | 73.09 | |
| CoVR-2Model Category=CVR Models2026.06 | 49.97 | 71.17 | |
| CoVR_EnrichModel Category=CVR Models2026.06 | 49.84 | 71.05 | |
| CoVRModel Category=CVR Models2026.06 | 48.53 | 70.25 | |
| LinCIRArch=ViT-G/14, Train=true2026.05 | 45.11 | 65.69 | |
| SADNModel Category=CIR Models2026.06 | 43.91 | 67.36 | |
| SEIZEArch=ViT-G/14, Train=false2026.05 | 43.05 | 65.85 | |
| IUDCModel Category=CIR Models2026.06 | 39.76 | 64.88 | |
| STiTchArch=ViT-G/14, Train=false2026.05 | 39.12 | 59.43 | |
| OSrCIRArch=ViT-G/14, Train=false2026.05 | 37.57 | 57.11 | |
| SEARLEArch=ViT-G/14, Train=true2026.05 | 34.81 | 55.71 | |
| CIReVLArch=ViT-G/14, Train=false2026.05 | 32.19 | 52.36 | |
| Pic2WordArch=ViT-G/14, Train=true2026.05 | 31.28 | 51.89 | |
| MagicLensBackbone=CLIP ViT-L/142026.07 | 30.7 | 52.5 | |
| FlowCIRBackbone=CLIP ViT-L/142026.07 | 29.7 | 48.9 | |
| Context-I2WBackbone=CLIP ViT-L/142026.07 | 27.8 | 48.9 | |
| MagicLensBackbone=CLIP ViT-B/322026.07 | 26.3 | 47.4 | |
| LinCIRBackbone=CLIP ViT-L/142026.07 | 26.3 | 46.5 | |
| SEARLEBackbone=CLIP ViT-L/142026.07 | 25.6 | 46.2 | |
| Pic2WordBackbone=CLIP ViT-L/142026.07 | 24.7 | 43.7 | |
| FlowCIRBackbone=CLIP ViT-B/322026.07 | 23.4 | 43.2 | |
| SEARLEBackbone=CLIP ViT-B/322026.07 | 22.9 | 42 | |
| PALAVRABackbone=CLIP ViT-B/322026.07 | 19.8 | 37.3 | |
| Text-onlyBackbone=CLIP ViT-B/322026.07 | 18 | 35.9 | |
| Text-onlyBackbone=CLIP ViT-L/142026.07 | 16.4 | 31.3 | |
| Image-onlyBackbone=CLIP ViT-L/142026.07 | 7.8 | 16.2 | |
| Image-onlyBackbone=CLIP ViT-B/322026.07 | 5.9 | 12.7 |