Image Generation on FFHQ 10k (train)
0.61SwAV-FIDVision-aided Adversarial training
Evaluation results
| Method | Links | |
|---|---|---|
| Vision-aided Adversarial trainingAugmentation Strategy=ADA, Fine-tuning=true2021.12 | 0.61 | |
| Vision-aided Adversarial trainingAugmentation Strategy=DiffAugment, Fine-tuning=true2021.12 | 0.73 | |
| ADAAugmentation Strategy=ADA2021.12 | 1.59 | |
| DiffAugmentAugmentation Strategy=DiffAugment2021.12 | 1.95 | |
| StyleGAN2Augmentation Strategy=none2021.12 | 3.85 |