Unconditional Image Generation on CelebA 128x128 (train test)
2.03FIDU-Net GAN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| U-Net GANFramework=PyTorch, Resolution=128x128, Training Iterations=800k2020.02 | 2.03 | 3.33 | |
| U-Net GANFramework=TensorFlow, Resolution=128x128, Training Iterations=800k2020.02 | 2.95 | 3.43 | |
| BigGANFramework=PyTorch, Resolution=128x128, Training Iterations=800k2020.02 | 3.7 | 3.08 | |
| BigGANFramework=TensorFlow, Resolution=128x128, Training Iterations=800k2020.02 | 4.54 | 3.23 | |
| COCO-GANFramework=TensorFlow, Resolution=128x1282020.02 | 5.74 | — | |
| PG-GANFramework=TensorFlow, Resolution=128x1282020.02 | 7.3 | — |