Metric: R@10 (higher is better)
| # | Model↕ | R@10▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | CoVR-BLIP-2 | 49.96 | No | CoVR-2: Automatic Data Construction for Composed... | 2023-08-28 | Code |
| 2 | CoLLM (finetuned - BLIP-L/16) | 39.1 | No | CoLLM: A Large Language Model for Composed Image... | 2025-03-25 | Code |
| 3 | SCOT (WACV 2025) | 38.45 | No | SCOT: Self-Supervised Contrastive Pretraining Fo... | 2025-01-12 | - |
| 4 | CoVR-BLIP-2 | 38.15 | No | CoVR-2: Automatic Data Construction for Composed... | 2023-08-28 | Code |
| 5 | MagicLens (CoCa L) | 38 | No | MagicLens: Self-Supervised Image Retrieval with ... | 2024-03-28 | Code |
| 6 | CoLLM (Pretrained - BLIP-L/16) | 34.6 | No | CoLLM: A Large Language Model for Composed Image... | 2025-03-25 | Code |
| 7 | ImageScope (CLIP-ViT-L/14) | 31.36 | No | ImageScope: Unifying Language-Guided Image Retri... | 2025-03-13 | Code |
| 8 | MagicLens (CLIP L) | 30.7 | No | MagicLens: Self-Supervised Image Retrieval with ... | 2024-03-28 | Code |
| 9 | CoLLM (Pretrained - CLIP-L/14) | 30.1 | No | CoLLM: A Large Language Model for Composed Image... | 2025-03-25 | Code |