https://github.com/stwind/gan-experiments

Various GAN experiments with Pytorch
https://github.com/stwind/gan-experiments

Last synced: about 1 year ago
JSON representation

Various GAN experiments with Pytorch

• Host: GitHub
• URL: https://github.com/stwind/gan-experiments
• Owner: stwind
• License: mit
• Created: 2020-12-27T16:17:00.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2021-03-18T14:23:18.000Z (about 5 years ago)
• Last Synced: 2025-03-24T21:51:08.959Z (about 1 year ago)
• Language: Jupyter Notebook
• Size: 87.9 MB
• Stars: 3
• Watchers: 3
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# GAN Experiments

Various GAN experiments with Pytorch.

## Contents

* [WGAN-GP On Toy Datasets](#wgan-gp-on-toy-datasets)

* [DCGAN with GP on MNIST](#dcgan-with-gp-on-mnist)

* [WGAN with Spectral Normalization on Flower102](#wgan-with-spectral-normalization-on-flower102)

## Notebooks

### [WGAN-GP On Toy Datasets](notebooks/wgan_gp_toy.ipynb)

Gaussian 8

![](images/wgan_gp_gaussian_8.png)

Gaussian 25

![](images/wgan_gp_gaussian_25.png)

Swiss Roll

![](images/wgan_gp_swiss_roll.png)

### [DCGAN with GP on MNIST](notebooks/wgan_gp_mnist.ipynb)

![](images/wgan_gp_mnist_stats.png)

![](images/wgan_gp_mnist_results.png)

### [WGAN with Spectral Normalization on Flower102](notebooks/wgan_specnorm_flower102.ipynb)

![](images/wgan_specnorm_flower102.png)

### [Stylegan2 Ada On Ukiyoe-faces](notebooks/stylegan2_ada_ukiyoe_faces.ipynb)

![](images/stylegan2_ukiyoue_faces_interp.png)

### [GANSpace PCA with Stylegan2 Ada On Ukiyoe-faces](notebooks/stylegan2_ganspace.ipynb)

![](images/stylegan2_ganspace_pca.png)

## References

* [[1704.00028] Improved Training of Wasserstein GANs](https://arxiv.org/abs/1704.00028)

* [[1801.04406] Which Training Methods for GANs do actually Converge?](https://arxiv.org/abs/1801.04406)

* [[1902.03984] Improving Generalization and Stability of Generative Adversarial Networks](https://arxiv.org/abs/1902.03984)

* [[1709.08894] On the regularization of Wasserstein GANs](https://arxiv.org/abs/1709.08894)

* [[1609.04468] Sampling Generative Networks](https://arxiv.org/abs/1609.04468)

  * [dribnet/plat](https://github.com/dribnet/plat)

* [[1511.08861] Loss Functions for Neural Networks for Image Processing](https://arxiv.org/abs/1511.08861)

* [[1807.00734] The relativistic discriminator: a key element missing from standard GAN](https://arxiv.org/abs/1807.00734)

* [[1801.03924] The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://arxiv.org/abs/1801.03924)

* [[1706.08500] GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium](https://arxiv.org/abs/1706.08500)

* [[1801.04406] Which Training Methods for GANs do actually Converge?](https://arxiv.org/abs/1801.04406)

* [[1512.09300] Autoencoding beyond pixels using a learned similarity metric](https://arxiv.org/abs/1512.09300)

* [[2002.04185] Smoothness and Stability in GANs](https://arxiv.org/abs/2002.04185)

* [[1705.09367] Stabilizing Training of Generative Adversarial Networks through Regularization](https://arxiv.org/abs/1705.09367)

* [[1802.05957] Spectral Normalization for Generative Adversarial Networks](https://arxiv.org/abs/1802.05957)

* [[2002.03754] Unsupervised Discovery of Interpretable Directions in the GAN Latent Space](https://arxiv.org/abs/2002.03754)

* [[2004.02546] GANSpace: Discovering Interpretable GAN Controls](https://arxiv.org/abs/2004.02546)