Text-to-Image Generation on Coco Prompt 6K
0.102rFIDFLUX VAE
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| FLUX VAETokenizer=FLUX VAE, CFG=Without, Resolution=256x256, Training steps=100K, Model parameters=2B2025.09 | 0.102 | 48.76 | 0.025 | 0.183 | 0.345 | -1.127 | 4.205 | 27.05 | 53.4 | |
| FLUX VAETokenizer=FLUX VAE, Resolution=256x256, Training Steps=100K, Number of Parameters=2B, CFG=True2025.09 | 0.102 | 35.78 | 0.018 | 0.242 | 0.397 | 0.162 | 5.411 | 31.21 | 77.5 | |
| AlignTokTokenizer=AlignTok, CFG=Without, Resolution=256x256, Training steps=100K, Model parameters=2B2025.09 | 0.443 | 33.46 | 0.014 | 0.21 | 0.655 | -0.581 | 4.963 | 28.87 | 65.8 | |
| AlignTokTokenizer=AlignTok, Resolution=256x256, Training Steps=100K, Number of Parameters=2B, CFG=True2025.09 | 0.443 | 30.27 | 0.015 | 0.249 | 0.603 | 0.564 | 5.573 | 32.21 | 84.9 |