Image-Text Retrieval on ShareGPT4V (Robustness Evaluation)
95.4Retrieval Score (Word Dropout p=0.3)HyFL-CLIP
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| HyFL-CLIP2026.07 | 95.4 | 95.44 | 96.45 | — | 97.37 | 56.52 | 63.59 | |
| HiMo-CLIPimplementation=evaluated using our own implementation2026.07 | 94.8 | 86.27 | 95.6 | — | 96.93 | 48.14 | 57.08 | |
| FineLIPimplementation=evaluated using our own implementation2026.07 | 93.74 | 85.48 | 94.8 | — | 96.23 | 45.34 | 53.42 | |
| Long-CLIPsource=obtained using checkpoints provided by the original authors2026.07 | 86.91 | 75.79 | 88.6 | — | 88.8 | 45.34 | 51.19 |