Partial implementation of ODE-GAN technique from the paper Training Generative Adversarial Networks by Solving Ordinary Differential Equations

Last update: Feb 10, 2022

Related tags

Deep Learning pytorch_odegan

Overview

ODE GAN (Prototype) in PyTorch

Partial implementation of ODE-GAN technique from the paper Training Generative Adversarial Networks by Solving Ordinary Differential Equations.

Caveat

This is not a faithful reproduction of the paper!

One of the many major difference is the use of gradient normalization to stabilize training (and avoid exploding gradients which lead to nans in generator + discriminator).
Another difference might be implementation of the regularization component.
Finally, this is a prototype to demonstrate the training regiment, without any focus for optimization of any kind - there's a lot of duplication of weights, caches etc throughout the code.

Training Regiment

By default, the model is trained on the CIFAR 10 dataset, with most of the parameters set in argparse.

Here is a tensorboard of a model being trained using RK2 (Heuns ODE step) for 250 epochs ~ 187500 update steps - Tensorboard Dev Log

Generated images

Training has not completed yet, here are images at the 60th epoch of training. Assuming nothing crashes in the next 200 epochs, there might be better results in later epochs.

You might also like...

NR-GAN: Noise Robust Generative Adversarial Networks

NR-GAN: Noise Robust Generative Adversarial Networks (CVPR 2020) This repository provides PyTorch implementation for noise robust GAN (NR-GAN). NR-GAN

59 Dec 11, 2022

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis Jungil Kong, Jaehyeon Kim, Jaekyoung Bae In our paper, we p

31 Dec 8, 2022

π-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis

π-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis Project Page | Paper | Data Eric Ryan Chan*, Marco Monteiro*, Pe

375 Dec 31, 2022

Flickr-Faces-HQ (FFHQ) is a high-quality image dataset of human faces, originally created as a benchmark for generative adversarial networks (GAN)

Flickr-Faces-HQ Dataset (FFHQ) Flickr-Faces-HQ (FFHQ) is a high-quality image dataset of human faces, originally created as a benchmark for generative

2.9k Dec 28, 2022

House-GAN++: Generative Adversarial Layout Refinement Network towards Intelligent Computational Agent for Professional Architects

House-GAN++ Code and instructions for our paper: House-GAN++: Generative Adversarial Layout Refinement Network towards Intelligent Computational Agent

122 Dec 28, 2022

A method that utilized Generative Adversarial Network (GAN) to interpret the black-box deep image classifier models by PyTorch.

3 Dec 29, 2022

Comments

Memory exploding

Hello First of all, thanks you for implementation of odegan.

When I trained a odegan using cifar10 dataset, I had memory exploding problem. So, I think this problem from version of the pytroch etc..

Can you tell me the version of the environment you used?

My environment is followed: pytorch 1.10.0 py3.9_cuda11.3_cudnn8.2.0_0 pytorch torchvision 0.11.1 py39_cu113 pytorch

Thanks

opened by Jinsung-Jeon 0

Partial implementation of ODE-GAN technique from the paper Training Generative Adversarial Networks by Solving Ordinary Differential Equations

Related tags

Overview

ODE GAN (Prototype) in PyTorch

Caveat

Training Regiment

Generated images

You might also like...

NR-GAN: Noise Robust Generative Adversarial Networks

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis

π-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis

Flickr-Faces-HQ (FFHQ) is a high-quality image dataset of human faces, originally created as a benchmark for generative adversarial networks (GAN)

Generate high quality pictures. GAN. Generative Adversarial Networks

Stable Neural ODE with Lyapunov-Stable Equilibrium Points for Defending Against Adversarial Attacks

Super-Fast-Adversarial-Training - A PyTorch Implementation code for developing super fast adversarial training

House-GAN++: Generative Adversarial Layout Refinement Network towards Intelligent Computational Agent for Professional Architects

A method that utilized Generative Adversarial Network (GAN) to interpret the black-box deep image classifier models by PyTorch.

Comments

Memory exploding

Owner

Somshubra Majumdar

Leibniz is a python package which provide facilities to express learnable partial differential equations with PyTorch

Deep learning library for solving differential equations and more

Simple Linear 2nd ODE Solver GUI - A 2nd constant coefficient linear ODE solver with simple GUI using euler's method

PyTorch implementation for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)

Official code for Score-Based Generative Modeling through Stochastic Differential Equations

Code for "Infinitely Deep Bayesian Neural Networks with Stochastic Differential Equations"

PyTorch implementation for SDEdit: Image Synthesis and Editing with Stochastic Differential Equations

PyTorch implementation for OCT-GAN Neural ODE-based Conditional Tabular GANs (WWW 2021)

Python framework for Stochastic Differential Equations modeling

FuseDream: Training-Free Text-to-Image Generationwith Improved CLIP+GAN Space OptimizationFuseDream: Training-Free Text-to-Image Generationwith Improved CLIP+GAN Space Optimization