Text-to-Image Synthesis on MS-COCO 2017 (test)
19.91FIDReg-VQ
Evaluation results
| Method | Links | |
|---|---|---|
| Reg-VQgenerative model=diffusion, resolution=256x2562023.03 | 19.91 | |
| PoE-GANResolution=256x256, Conditioning=Text2021.12 | 20.5 | |
| Gumbel-VQgenerative model=diffusion, resolution=256x2562023.03 | 22.21 | |
| VQ-GANgenerative model=diffusion, resolution=256x2562023.03 | 23.75 | |
| DM-GAN + CLResolution=256x256, Conditioning=Text2021.12 | 29.9 | |
| VQ-VAEgenerative model=diffusion, resolution=256x2562023.03 | 38.84 | |
| DF-GANResolution=256x256, Conditioning=Text2021.12 | 45.2 |