paper: (https://arxiv.org/pdf/2102.12593.pdf)
This code seems to understand the structure & texture information provided by training images, but doesn't produce meaningful results yet. I will keep debugging and posting the results here.
- AST block
- FST blocks
- Style encoder
- Content encoder
- Double branch discriminator
- POLIN
- AdaPOLIN
- adv loss, reconstruction loss, feature matching loss (hook on activations)