DiffAugment-BigGAN

Reported on 5 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical5 results

Image GenerationonCIFAR-10 (10% data)
FID· 2020-06-18
22.4
best: 14.5 (DiffAugment-StyleGAN2)
SOTA
Differentiable Augmentation for Data-Efficient GAN Training arXiv:2006.10738
Image GenerationonCIFAR-10 (20% data)
FID· 2020-06-18
14.04
best: 12.15 (DiffAugment-StyleGAN2)
SOTA
Differentiable Augmentation for Data-Efficient GAN Training arXiv:2006.10738
Image GenerationonCIFAR-10
FID· 2020-06-18
4.61
best: 1.22 (GMem)
Differentiable Augmentation for Data-Efficient GAN Training arXiv:2006.10738
Image GenerationonImageNet 128x128
FID· 2020-06-18
6.8
best: 1.26 (SiD2)
Differentiable Augmentation for Data-Efficient GAN Training arXiv:2006.10738
Image GenerationonImageNet 128x128
IS· 2020-06-18
100.8
best: 190.5 (VDM++)
Differentiable Augmentation for Data-Efficient GAN Training arXiv:2006.10738