Composed Image Retrieval on FBCIR-Data
41.6Rs@1BGE-base
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BGE-baseModel Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 41.6 | 10 | |
| RzenEmbed-7BModel Category=VLM-based Models, Evaluation Protocol=Pretrained2026.03 | 39.3 | 43 | |
| BGE-largeModel Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 38.8 | 2 | |
| CLIP4CIR-RN50Model Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 33.4 | 53 | |
| GME-7BModel Category=VLM-based Models, Evaluation Protocol=Pretrained2026.03 | 32.9 | 55 | |
| CLIP4CIR-RN50x4Model Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 31.2 | 65 | |
| SEARLE-largeModel Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 30.2 | 70 | |
| GME-2BModel Category=VLM-based Models, Evaluation Protocol=Pretrained2026.03 | 29.7 | 54 | |
| SEARLE-baseModel Category=CLIP-based Models, Evaluation Protocol=Pretrained2026.03 | 29.3 | 27 | |
| MM-Embed-7BModel Category=VLM-based Models, Evaluation Protocol=Pretrained2026.03 | 26.7 | 60 |