Composed Image Retrieval on Fashion-IQ Dress (val)
38.57Recall@10WISER
Evaluation results
| Method | Links | ||
|---|---|---|---|
| WISERBackbone=ViT-G/14, Training-free=true2026.02 | 38.57 | 58.35 | |
| WISERBackbone=ViT-L/14, Training-free=true2026.02 | 38.42 | 55.92 | |
| LinCIRBackbone=ViT-G/14, Training-free=false2026.02 | 38.08 | 60.88 | |
| WISERBackbone=ViT-B/32, Training-free=true2026.02 | 37.48 | 56.97 | |
| CoTMRBackbone=ViT-G/14, Training-free=true2026.02 | 34.51 | 57.36 | |
| OSrCIRBackbone=ViT-G/14, Training-free=true2026.02 | 33.02 | 54.78 | |
| CoTMRBackbone=ViT-L/14, Training-free=true2026.02 | 31.18 | 55.04 | |
| CoTMRBackbone=ViT-B/32, Training-free=true2026.02 | 31.09 | 54.54 | |
| OSrCIRBackbone=ViT-L/14, Training-free=true2026.02 | 29.7 | 51.81 | |
| OSrCIRBackbone=ViT-B/32, Training-free=true2026.02 | 29.35 | 50.37 | |
| CIReVLBackbone=ViT-G/14, Training-free=true2026.02 | 27.07 | 49.53 | |
| AutoCIRBackbone=ViT-B/32, Training-free=true2026.02 | 26.52 | 46.36 | |
| AutoCIRBackbone=ViT-G/14, Training-free=true2026.02 | 26.18 | 47.69 | |
| LDREBackbone=ViT-G/14, Training-free=true2026.02 | 26.11 | 51.12 | |
| HITBackbone=ViT-L/14, Training-free=false2026.02 | 25.6 | 47.1 | |
| CIReVLBackbone=ViT-B/32, Training-free=true2026.02 | 25.29 | 46.36 | |
| MOABackbone=ViT-L/14, Training-free=false2026.02 | 25.2 | 48.5 | |
| AutoCIRBackbone=ViT-L/14, Training-free=true2026.02 | 24.94 | 45.81 | |
| CIReVLBackbone=ViT-L/14, Training-free=true2026.02 | 24.79 | 44.76 | |
| LDREBackbone=ViT-L/14, Training-free=true2026.02 | 22.93 | 46.76 | |
| LinCIRBackbone=ViT-L/14, Training-free=false2026.02 | 20.92 | 42.44 | |
| SEARLEBackbone=ViT-L/14, Training-free=false2026.02 | 20.48 | 43.13 | |
| LDREBackbone=ViT-B/32, Training-free=true2026.02 | 19.97 | 41.84 | |
| SEARLEBackbone=ViT-B/32, Training-free=false2026.02 | 18.54 | 39.51 | |
| PALAVRABackbone=ViT-B/32, Training-free=false2026.02 | 17.25 | 35.94 |