Composed Image Retrieval on GeneCIS 1.0 (test)
21.1Recall@1MMRet-MLLM
Evaluation Results
| Method | Links | |
|---|---|---|
| MMRet-MLLMBackbone=LLaVA-1.6, # Params=7.57B, Zero-shot evaluation=true2024.12 | 21.1 | |
| MMRet-LargeBackbone=CLIP-L, # Params=428M, Zero-shot evaluation=true2024.12 | 18.1 | |
| MMRet-BaseBackbone=CLIP-B, # Params=149M, Zero-shot evaluation=true2024.12 | 18 | |
| MagicLens-B*Backbone=CoCa-B, # Params=267M, Zero-shot evaluation=true2024.12 | 17.4 | |
| CIREVLBackbone=CLIP-G, # Params=14.6B+, Zero-shot evaluation=true2024.12 | 17.4 | |
| MagicLens-L*Backbone=CoCa-L, # Params=613M, Zero-shot evaluation=true2024.12 | 16.7 | |
| MagicLens-LBackbone=CLIP-L, # Params=465M, Zero-shot evaluation=true2024.12 | 16.3 | |
| CIREVLBackbone=CLIP-B, # Params=12.3B+, Zero-shot evaluation=true2024.12 | 15.9 | |
| CIREVLBackbone=CLIP-L, # Params=12.5B+, Zero-shot evaluation=true2024.12 | 15.9 | |
| MagicLens-BBackbone=CLIP-B, # Params=166M, Zero-shot evaluation=true2024.12 | 15 | |
| CompoDiffBackbone=CLIP-L, # Params=568M, Zero-shot evaluation=true2024.12 | 14.9 | |
| SEARLEBackbone=CLIP-L, # Params=442M, Zero-shot evaluation=true2024.12 | 12.3 | |
| Pic2WordBackbone=CLIP-L, # Params=429M, Zero-shot evaluation=true2024.12 | 11.2 |