Image-to-Text Retrieval on ShareGPT4V (unseen)
95.8R@1Long-CLIP-L
Evaluation Results
| Method | Links | |
|---|---|---|
| Long-CLIP-LEvaluation Protocol=q_i -> c_t2026.06 | 95.8 | |
| EVA-CLIP-18BEvaluation Protocol=q_i -> c_t2026.06 | 94 | |
| EVA-CLIP-8BEvaluation Protocol=q_i -> c_t2026.06 | 93.1 | |
| ELVA-7BEvaluation Protocol=q_i -> c_t2026.06 | 92 | |
| LamRA-Ret-7BEvaluation Protocol=q_i -> c_t2026.06 | 88.1 | |
| UniIR-CLIPEvaluation Protocol=q_i -> c_t2026.06 | 84.1 | |
| E5-VEvaluation Protocol=q_i -> c_t2026.06 | 84 | |
| CLIP-LEvaluation Protocol=q_i -> c_t2026.06 | 81.8 | |
| MagicLens-LEvaluation Protocol=q_i -> c_t2026.06 | 60.9 |