torchgan

The torchgan package consists of various generative adversarial networks and utilities that have been found useful in training them. This package provides an easy to use API which can be used to train popular GANs as well as develop newer variants. The core idea behind this project is to facilitate easy and rapid generative adversarial model research.

GETTING STARTED

• Installation
  • Pip Installation
  • Conda Installation
  • Install from Source
• Supporting and Citing
• Dependencies
  • Mandatory Dependencies
  • Optional Dependencies
• Philosophy
• Contributing
  • Contribution Guidelines
  • Contributors
• Starter Example
• License

API DOCUMENTATION

• torchgan.layers
  • Residual Blocks
  • Densenet Blocks
  • Self Attention
  • Spectral Normalization
  • Minibatch Discrimination
  • Virtual Batch Normalization
• torchgan.logging
  • Backends
  • Logger
  • Visualization
• torchgan.losses
  • Loss
  • Least Squares Loss
  • Minimax Loss
  • Boundary Equilibrium Loss
  • Energy Based Loss
  • Wasserstein Loss
  • Mutual Information Penalty
  • Dragan Loss
  • Auxillary Classifier Loss
  • Feature Matching Loss
  • Historical Averaging
• torchgan.metrics
  • Metric
  • Classifier Score
• torchgan.models
  • Vanilla GAN
  • Deep Convolutional GAN
  • Conditional GAN
  • InfoGAN
  • AutoEncoders
  • Auxiliary Classifier GAN
• torchgan.trainer
  • Base Trainer
  • Trainer
  • Parallel Trainer