Text-to-Image Alignment Evaluation on COCO LAFITE samples
87.3Kendall Tau_cMID
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MIDBackbone=CLIP (ViT-L/14)2022.05 | 87.3 | 75.6 | |
| MIDBackbone=CLIP (ViT-B/32)2022.05 | 74.6 | 64.6 | |
| OFA-Captioning+CLIP-SBackbone=OFA-Large + CLIP (ViT-L/14)2022.05 | 73.7 | 63.8 | |
| OFA-Captioning+CLIP-SBackbone=OFA-Large + CLIP (ViT-B/32)2022.05 | 72 | 62.3 | |
| CLIP-R-PrecisionBackbone=CLIP (ViT-L/14)2022.05 | 69.6 | 58.1 | |
| CLIP-R-PrecisionBackbone=CLIP (ViT-B/32)2022.05 | 66 | 56.1 | |
| InfoNCEBackbone=CLIP (ViT-L/14)2022.05 | 64.8 | 56.1 | |
| CLIP-SBackbone=CLIP (ViT-L/14)2022.05 | 52.2 | 45.2 | |
| SOABackbone=YOLO-V32022.05 | 47.3 | 51.8 | |
| InfoNCEBackbone=CLIP (ViT-B/32)2022.05 | 44.1 | 38.2 | |
| CLIP-SBackbone=CLIP (ViT-B/32)2022.05 | 40.8 | 35.3 |