Text-to-image retrieval on MSCOCO 48 (test)
52.5R@1COSMOS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| COSMOSPre-training Data=Merged-30M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 52.5 | 77.2 | 84.9 | |
| COSMOSPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 50.9 | 76 | 84.2 | |
| SigLIPPre-training Data=Merged-30M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 49 | 74.5 | 83.2 | |
| COSMOSPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 48.9 | 74.3 | 82.9 | |
| CLIPPre-training Data=Merged-30M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 48.2 | 73.9 | 82.4 | |
| SigLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 46.5 | 72.8 | 81.9 | |
| LiTPre-training Data=2.5B, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 45.6 | 70.8 | 79.7 | |
| DreamLIPPre-training Data=Merged-30M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 44.9 | 69.9 | 79.2 | |
| CLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 44.7 | 71.2 | 80.3 | |
| DreamLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 42.4 | 68.5 | 78.2 | |
| OpenCLIPPre-training Data=LAION-2B, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 41.7 | 67.1 | 76.5 | |
| MetaCLIPPre-training Data=2.5B, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 41.4 | 67.2 | 77 | |
| DreamLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 40.7 | 66.5 | 76.4 | |
| SigLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 40.4 | 67 | 76.6 | |
| COSMOSPre-training Data=CC3M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 40.1 | 66.8 | 77.1 | |
| CLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 39.8 | 66.4 | 76.6 | |
| OpenCLIPPre-training Data=DataComp-1B, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 39.8 | 65.5 | 75 | |
| OpenCLIPPre-training Data=LAION-400M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 37.9 | 63.2 | 73.3 | |
| LaSF-CLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 31.4 | 57.3 | 68.1 | |
| DreamLIPPre-training Data=CC3M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 29.8 | 55.4 | 66.3 | |
| SigLIPPre-training Data=CC3M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 28.3 | 54.4 | 65.9 | |
| CLIPPre-training Data=CC3M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 27.2 | 53.5 | 65 | |
| SLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 26.8 | 51.4 | 62.7 | |
| MLLM-APre-training Data=CC3M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 26.5 | 51.1 | 62.7 | |
| LaCLIPPre-training Data=CC12M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 26.5 | 51.2 | 62.6 | |
| MaskCLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 25.5 | 49.7 | 61.3 | |
| SF-CLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 25.5 | 50.2 | 61.3 | |
| SLIPPre-training Data=YFCC15M, Evaluation Protocol=Zero-shot, Vision Encoder Architecture=ViT-B/162024.12 | 21.5 | 44.4 | 56.3 |