StudioGAN is Pohang University CVLab Min-guk KangPostech StudioGAN is an open source library developed by Min-Kuk Kang of CVLab, Postech. It is based on PyTorch and various GAN algorithms are implemented. The StudioGAN contains many major algorithm implementations such as DCGAN, LSGAN, and WGAN, so it is considered to be very useful for research purposes such as mutual performance comparison.
According to the code description, it is said that each algorithm is optimized in terms of memory requirements and complexity than the original implementation, and a pre-trained model is also disclosed. Besides, the latest technologies such as multi-GPU, mixed precision, and synchronized batch normalization are also reflected, so I think it is one of the best starting points for GAN related research. Here is a link to StudioGAN's github repository.
