A Convolutional block.

Multi-scale (scale) discriminator

Merge all discriminator related updates into single model for better multi-gpu usage

HDTF dataset class

VAE (Variational Autoencoder) class for image processing.

VFHQ dataset class

Band-limited downsampling, for better preservation of the input signal.

Discriminator similar to Pix2Pix

Simple block for processing video (encoder).

Convert any spelled-out numbers into arabic numbers, while handling: - remove any commas - keep the suffixes such as: `1960s`, `274th`,

Applies British-American spelling mappings as listed in [1]. [1] https://www.tysto.com/uk-us-spelling-list.html

Create image pyramide for computing pyramide perceptual loss.

Select the sample with the highest log probabilities, penalized using either a simple length normalization or Google NMT paper's length penal

A thin wrapper around `GPT2TokenizerFast` providing quick access to special tokens

Dataset class for loading and processing video data Each video can be represented as: - Concatenated frame images - '.mp4' or '.gif'

An abstract class representing a face detector. Any other face detection implementation must subclass it. All subclasses must implement ``det

Enum class defining the type of landmarks to detect. ``_2D`` - the detected points ``(x,y)`` are detected in a 2D space and follow the visible co

Vgg19 network for perceptual loss.