| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| UNIVL-ImgGen single and multiple masks | UNIVL + T5 | FID10.49 | 10 | 12d ago | |
| VisDial 1.0 (test) | GILL | CLIP Similarity0.645 | 9 | 3mo ago | |
| VIST 28 (test) | GILL | CLIP Similarity0.641 | 7 | 3mo ago | |
| COCO 2017 (val) | UniVL | FID16.99 | 3 | 12d ago |