Skip to content

Tanmoy0077/GAN-Implementations-in-PyTorch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GAN Implementations in PyTorch

Description

This repository contains PyTorch implementations of three Generative Adversarial Networks (GAN) architectures:

  1. CycleGAN: Cycle-Consistent Adversarial Networks for unpaired image-to-image translation.
  2. WGAN: Wasserstein GAN for improved training stability and more interpretable loss metric.
  3. VanillaGAN: The original GAN architecture for generating data from random noise.

CycleGAN

CycleGAN allows for image-to-image translation tasks where paired examples are not available. The key idea is to train two generator-discriminator pairs to perform mutual mappings between two domains.

WGAN

Wasserstein GAN improves the training of GANs by using the Earth Mover's distance (Wasserstein distance) as a loss metric. This approach helps to address issues of mode collapse and provides more meaningful loss values.

VanillaGAN

The VanillaGAN is the original formulation of GANs, where a generator network learns to produce data resembling the training data, and a discriminator network learns to distinguish between real and generated data.

Dataset

The datasets used for training these GANs can be found at the following links:

Rest are mentioned in the jupyter notebooks

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Acknowledgements


For further information and questions, feel free to open an issue or contact the repository maintainers.

About

This repository contains PyTorch implementations of three Generative Adversarial Networks (GAN) architectures.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published