Skip to content

shyam671/Multi-Agent-Diverse-Generative-Adversarial-Networks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

29 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multi Agent Diverse Generative Adversarial Network (MAD-GAN)

Pytorch Code for MAD-GAN.

Requirements

  • pytorch >=0.4.0
  • torchvision ==0.2.0
  • Jupyter Notebook

Datset

  • A distribution of 1D GMM having five mixture components with modes at 10, 20, 60, 80 and 110, and standard deviations of 3, 3, 2, 2 and 1, respectively.

Usage

  • Run the Simple_GANs.ipynb to generate the results for vanilla GAN.
  • For MAD-GAN run the Mad_GANs.ipynb to compare the results of vanillaGAN over MAD-GAN.
  • Number of generater of MAD-GAN can by changing 'num_gen' and G.params().

Results

Drag RacingDrag Racing

             [Left]: Result of VanillaGAN                [Right]: Result of MAD-GAN (number of generator = 4)

Resources

Acknowledgements

  • https://github.com/wiseodd
  • Ghosh, Arnab, et al. "Multi-agent diverse generative adversarial networks." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018

About

Easy-to-follow Pytorch tutorial Notebook for Multi-Agent-Diverse-Generative-Adversarial-Networks

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published