implement GANs and VAE using pytorch
GANs include DCGAN LSGAN WGAN WGAN-GP CGAN(use DCGAN architecture) VAE is from the paper Auto-Encoding Variational Bayes
in my code, i use MNIST to test the network, but its very easy to switch dataset of you own
here is result when i training DCGAN at 36 epoch
second are some humanface generate by WGAN-GP
the image use to train the WGAN is from baidu image
next is the gif from training a VAE use MNIST datasset
leftside is the reconstruction image,middle is the training image, and rightside is the image generate from the noise
next is the result from training a CDCGAN use MNIST dataset, each row represent a class of digit