Multimodal Retrieval (Image, Text) on OVEN
75.3R@5V-Retrver-7B*
Evaluation results
| Method | Links | ||
|---|---|---|---|
| V-Retrver-7B*Evaluation Protocol=Zero-shot, Training Condition=Held-out tasks (no exposure)2026.02 | 75.3 | 61.1 | |
| CLIP-SFRetrieval Strategy=Embedding-based, Fusion Strategy=Score-level Fusion, Re-ranking=false, Retrieval Pool=Task-specific2025.03 | 67.6 | — | |
| UniIR-CLIP_SFEvaluation Protocol=Supervised2026.02 | 67.6 | 44.4 | |
| LamRA-7B*Evaluation Protocol=Zero-shot, Training Condition=Held-out tasks (no exposure)2026.02 | 65.1 | 50.9 | |
| CLIP-SFApproach=Embedding-based, Fusion=SF2025.03 | 60.2 | — | |
| BLIP-FFApproach=Embedding-based, Fusion=FF2025.03 | 57.8 | — | |
| Vision-R1-7BEvaluation Protocol=Zero-shot2026.02 | 57.4 | 41.9 | |
| BLIP-FFRetrieval Strategy=Embedding-based, Fusion Strategy=Feature-level Fusion, Re-ranking=false, Retrieval Pool=Task-specific2025.03 | 55.8 | — | |
| UniIR-BLIP_FFEvaluation Protocol=Supervised2026.02 | 55.8 | 37 | |
| Qwen2.5-VL-7BEvaluation Protocol=Zero-shot2026.02 | 53.6 | 39.5 | |
| GENIUSRetrieval Strategy=Generative, Fusion Strategy=None, Re-ranking=true, Retrieval Pool=Task-specific2025.03 | 52.5 | — | |
| GENIUSApproach=Generative, Re-ranking=true2025.03 | 48.6 | — | |
| GENIUSApproach=Generative2025.03 | 36.5 | — | |
| GENIUSRetrieval Strategy=Generative, Fusion Strategy=None, Re-ranking=false, Retrieval Pool=Task-specific2025.03 | 36.4 | — |