Text-to-Image Generation on HPSv2 Prompt
0.102rFIDFLUX VAE
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| FLUX VAETokenizer=FLUX VAE, CFG=true, Resolution=256x256, Training steps=100K2025.09 | 0.102 | 0.224 | 0.373 | 0.09 | 5.478 | 31.59 | 73.7 | |
| FLUX VAETokenizer=FLUX VAE, CFG=Without, Resolution=256x256, Training steps=100K, Model parameters=2B2025.09 | 0.102 | 0.17 | 0.372 | -1.206 | 4.142 | 26.59 | 56.7 | |
| AlignTokTokenizer=AlignTok, CFG=true, Resolution=256x256, Training steps=100K2025.09 | 0.443 | 0.231 | 0.626 | 0.366 | 5.604 | 33.12 | 79.2 | |
| AlignTokTokenizer=AlignTok, CFG=Without, Resolution=256x256, Training steps=100K, Model parameters=2B2025.09 | 0.443 | 0.192 | 0.628 | -0.764 | 4.955 | 28.2 | 64.5 |