Image-Text Retrieval on MS-COCO (test)
0.8532Rt@1BLIP-2
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| BLIP-22024.05 | 0.8532 | 0.6826 | 507.46 | 528.63 | 678.45 | 900.77 | 752 | 1,612 | |
| PITOMEbase_model=BLIP-2, r=0.975, training=without training2024.05 | 0.8412 | 0.6737 | 504.95 | 527.06 | 434.5 | 564.78 | 513 | 924 | |
| PITOMEbase_model=BLIP-2, r=0.95, training=without training2024.05 | 0.8229 | 0.6554 | 494.92 | 518.44 | 296.93 | 390.77 | 333 | 634 | |
| BLIP2024.05 | 0.8182 | 0.6436 | 494.34 | 516.03 | 55.14 | 65.54 | 13 | 710 | |
| PITOMEbase_model=BLIP, r=0.95, training=with training2024.05 | 0.8044 | 0.6391 | 493.33 | 512.66 | 38.85 | 47.65 | 51 | 430 | |
| PITOMEbase_model=BLIP, r=0.95, training=without training2024.05 | 0.7946 | 0.625 | 485.99 | 506.65 | 38.85 | 47.65 | 51 | 430 | |
| ALBEF2024.05 | 0.7694 | 0.6024 | 478.39 | 500.44 | 55.14 | 65.54 | 43 | 529 | |
| METER2024.05 | 0.7616 | 0.5708 | — | 495.95 | — | — | — | — | |
| CLIP-L2024.05 | 0.7078 | 0.5379 | 478.18 | — | 80.85 | — | 210 | — | |
| ViLT2024.05 | 0.615 | 0.427 | 420.2 | 439.2 | — | 55.9 | — | — |