Skip to content

(CVPR 2021) DC-VAE: Dual Contradistinctive Generative Autoencoder

License

Notifications You must be signed in to change notification settings

mlpc-ucsd/DC-VAE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Dual Contradistinctive Generative AutoEncoder (CVPR 2021)

Project | Paper | Usage | Citation

A generative autoencoder model with dual contradistinctive losses to improve generative autoencoder that performs simultaneous inference (reconstruction) and synthesis (sampling). Our model, named dual contradistinctive generative autoencoder (DC-VAE), integrates an instance-level discriminative loss (maintaining the instance-level fidelity for the reconstruction/synthesis) with a set-level adversarial loss (encouraging the set-level fidelity for there construction/synthesis), both being contradistinctive.


Usage

  • Clone the repository

    git clone https://github.com/mlpc-ucsd/DC-VAE
    cd DC-VAE
    
  • Setup the conda environment

    conda env create -f dcvae_env.yml
    conda activate dcvae_env
    
  • Train the network on CIFAR-10

    python train.py
    

Citation

If you find this project useful for your research, please cite the following work.

@InProceedings{Parmar_2021_CVPR,
    author    = {Parmar, Gaurav and Li, Dacheng and Lee, Kwonjoon and Tu, Zhuowen},
    title     = {Dual Contradistinctive Generative Autoencoder},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2021},
    pages     = {823-832}
}

Credits

We found the following libraries helpful in our research.

  • FID - computing the FID score
  • IS - computing the Inception Score.
  • AutoGAN - model architecture for the low resolution experiments experiments
  • ProGAN - model architecture for the high resolution experiments.

Acknowledgements

This work is funded by NSF IIS- 1717431 and NSF IIS-1618477. We thank Qualcomm Inc. for an award support. The work was performed when G. Parmar and D. Li were with UC San Diego.


About

(CVPR 2021) DC-VAE: Dual Contradistinctive Generative Autoencoder

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages