TorchGAN: A Flexible Framework for GAN Training and Evaluation

TL;DR

TorchGAN is a flexible, extensible PyTorch framework that simplifies GAN training and evaluation, supporting various models and metrics with minimal overhead, thus facilitating research and development in generative modeling.

Contribution

It introduces a modular, easy-to-use framework for GANs that supports customization, multiple models, and evaluation metrics, with benchmarking showing negligible overhead.

Findings

Supports numerous GAN models and metrics

Demonstrates ease of customization and extensibility

Achieves near-zero overhead compared to vanilla PyTorch

Abstract

TorchGAN is a PyTorch based framework for writing succinct and comprehensible code for training and evaluation of Generative Adversarial Networks. The framework's modular design allows effortless customization of the model architecture, loss functions, training paradigms, and evaluation metrics. The key features of TorchGAN are its extensibility, built-in support for a large number of popular models, losses and evaluation metrics, and zero overhead compared to vanilla PyTorch. By using the framework to implement several popular GAN models, we demonstrate its extensibility and ease of use. We also benchmark the training time of our framework for said models against the corresponding baseline PyTorch implementations and observe that TorchGAN's features bear almost zero overhead.