Text-to-image retrieval on CLEVR-DS-S (test)
14.66Recall@1DetailCLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DetailCLIPSingle Feature=False, Input=CC@102022.08 | 14.66 | 32.77 | 32.78 | |
| CLIPSingle Feature=False, Input=CC@102022.08 | 14.57 | 32.55 | 37.12 | |
| CLIPSingle Feature=True, Input=Full Image2022.08 | 4.43 | 16.25 | 30.32 |