Text-to-Image Retrieval on MS-COCO Easy
0.676Recall@1BEIT-3
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BEIT-3Config=COCO-Finetuned, Zero-Shot=false2026.02 | 0.676 | 0.959 | 0.973 | |
| CLIPBackbone=ViT-H, QuickGELU=true, Zero-Shot=true2026.02 | 0.649 | 0.919 | 0.973 | |
| Visual Model Checking + CLIPZero-Shot=true2026.02 | 0.592 | 0.959 | 0.98 | |
| SigLIPParameters=400M, Zero-Shot=true2026.02 | 0.568 | 0.811 | 0.892 | |
| Visual Model Checking + BEITZero-Shot=false2026.02 | 0.551 | 0.939 | 0.98 | |
| CLIPBackbone=ViT-B, Zero-Shot=true2026.02 | 0.486 | 0.85 | 0.892 | |
| Visual Model Checking + ALIGNZero-Shot=true2026.02 | 0.469 | 0.857 | 0.878 | |
| ALIGNZero-Shot=true2026.02 | 0.459 | 0.469 | 0.838 | |
| Visual Model Checking (base)Zero-Shot=true2026.02 | 0.432 | 0.55 | 0.784 |