Simple implementation of Analyzing and Improving the Image Quality of StyleGAN (http://arxiv.org/abs/1912.04958).
The project is dedicated to the rapid prototyping of StyleGAN2-like networks using the building blocks defined here. It includes
- modulated conv layers
- minibatch std layer
- mapping network
- noise with learnable magnitude
- standard blocks for generator and discriminator
- equalized lr
- scaled activations
- losses
- projection method
If you need a complete network as in the paper you can easily implement it knowing all the parameters and dimensions. See the toy example here.
I would appreciate help with testing.
- Base classes for Generator and Discriminator
- Arrange the training procedure.