r""" An empty class used for configuration members

Composes several transforms together. Args: transforms (list of ``Transform`` objects): list of transforms to compose. Example:

r""" Implementation of NVIDIA's progressive GAN.

r""" Class implementing all tools necessary for a GAN to take into account class conditionning while generating a model (cf Odena's AC-GAN)

r""" A class managing a progressive GAN training. Logs, chekpoints, visualization, and number iterations are managed here.

r"""Abstract class: the basic framework for GAN training.

r""" Loss criterion class. Must define 4 members: sizeDecisionLayer : size of the decision layer of the discrimator getCriterion : how th

r""" A handy refactor that allows the user to: - initialize one layer's bias to zero - apply He's initialization at runtime

r""" Cross entropy loss.

This class manages all datasets and given the config for each of them it returns the corresponding DataLoader object and

r""" A class managing a progressive GAN training. Logs, chekpoints, visualization, and number iterations are managed here.

r""" "Which training method of GANs actually converge" https://arxiv.org/pdf/1801.04406.pdf

r""" Mean Square error loss.

r""" Paper WGANGP loss : linear activation for the generator. https://arxiv.org/pdf/1704.00028.pdf