Code used for TransGAN: Two Pure Transformers Can Make One Strong GAN, and That Can Scale Up.

• checkpoint gradient using torch.utils.checkpoint
• 16bit precision training
• Distributed Training (Faster!)
• IS/FID Evaluation
• Gradient Accumulation
• Stronger Data Augmentation
• Self-Modulation

python exp/cifar_train.py

First download the cifar checkpoint and put it on ./cifar_checkpoint. Then run the following script.

python exp/cifar_test.py

README waits for updated

Codebase from AutoGAN, pytorch-image-models

if you find this repo is helpful, please cite

@article{jiang2021transgan,
  title={TransGAN: Two Transformers Can Make One Strong GAN},
  author={Jiang, Yifan and Chang, Shiyu and Wang, Zhangyang},
  journal={arXiv preprint arXiv:2102.07074},
  year={2021}
}