Image Generation on MetFaces (train)
24.77FIDSQ-GAN + ADA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SQ-GAN + ADABase=StyleGAN2, Augmentation=ADA2025.03 | 24.77 | 4.93 | 0.008 | |
| StyleGAN2 + ADAAugmentation=ADA2025.03 | 28.1 | 4.29 | 0.006 | |
| StyleGAN2 + CR + ADARegularization=CR, Augmentation=ADA2025.03 | 29.91 | 4.2 | 0.008 | |
| SQ-GAN + CBIBase=StyleGAN2, Initialization=Knowledge-enhanced codebook initialization2025.03 | 35.44 | 4.02 | 0.022 | |
| SQ-GANBase=StyleGAN22025.03 | 41.12 | 3.31 | 0.024 | |
| StyleGAN2 + CRRegularization=CR2025.03 | 48.89 | 3.24 | 0.029 | |
| StyleGAN22025.03 | 53.21 | 3.16 | 0.035 | |
| StyleGAN2 + DigGANRegularization=DigGAN2025.03 | 53.97 | 3.09 | 0.031 | |
| StyleGAN2 + KD-DLGANRegularization=KD-DLGAN2025.03 | 54.05 | 3.11 | 0.029 | |
| StyleGAN2 + LeCamRegularization=LeCam2025.03 | 56.67 | 2.43 | 0.102 |