Image-to-Context (I-T) Retrieval on (test)
68.92R@1CP-CLIP
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CP-CLIPBackbone=ViT-B/16, Model Variant=(D)2026.06 | 68.92 | 87.77 | 92.14 | 95.56 | 98.55 | 77.16 | |
| CP-CLIPBackbone=ViT-L/16, Model Variant=(D)2026.06 | 64.77 | 85.02 | 90.13 | 93.83 | 97.52 | 73.51 | |
| CP-CLIPBackbone=ViT-B/16, Model Variant=(fp)2026.06 | 64.2 | 86.55 | 91.99 | 95.73 | 98.75 | 73.85 | |
| CLIPBackbone=ViT-B/162026.06 | 58.55 | 80.67 | 86.54 | 91.71 | 96.85 | 68.2 | |
| SigLIPBackbone=ViT-B/162026.06 | 38.65 | 56.15 | 63.86 | 71.56 | 81.92 | 47 | |
| SigLIPBackbone=ViT-B/16, Model Variant=(D)2026.06 | 20.71 | 39.87 | 48.83 | 59.65 | 74.33 | 30.15 |