Linear layer with equalized learning rate and custom learning rate multiplier.

A single linear layer which does a 1x1 conv

Pretrained InceptionV3 network returning feature maps

InceptionC block patched for FID computation

Things to do at the end of each layer.

Conv layer with equalized learning rate and custom learning rate multiplier.

InceptionA block patched for FID computation

Adds static noise to output Args: noise (th.tensor): Noise to be added

Adapted from https://github.com/foamliu/Autoencoder See LICENSE_AUTOENCODER

r""" A PyTorch implementation of the lookahead wrapper for GANs. This optimizer performs the lookahead step on both the discriminator and gen

takes 2 distances, puts through fc layers, spits out value between [0,1] (if use_sigmoid is True)

First InceptionE block patched for FID computation

Second InceptionE block patched for FID computation

Adapted from https://github.com/koshian2/inception-vae

Adapted from https://github.com/AntixK/PyTorch-VAE See LICENSE_AUTOENCODER

adds noise. noise is per pixel (constant over channels) with per-channel weight

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG

Adapted from https://github.com/NVIDIA/pix2pixHD See LICENSE-VGG