• SSD-GAN
• • • Dependencies
    • Training
    • Testing
    • Results
    • Bibtex
    • Acknowledgement
    • Contact

SSD-GAN

Pytorch implementation of our paper: "SSD-GAN: Measuring the Realness in the Spatial and Spectral Domains".

Dependencies

pip install torch-mimicry

Training

• Training SSD-SNGAN on CIFAR100

python train_cifar100.py

• Training SSD-SNGAN on STL10

python train_stl10.py

• Training SSD-SNGAN on LSUN-Bedroom

python train_lsun.py

• Tensorboard visualizations

tensorboard --logdir=./log/YOUR_LOGDIR

Testing

python test --dataset cifar100 --log_dir ./log/cifar100

Results

FID scores.

Generations of SSD-SNGAN trained on LSUN-bedroom at 128×128.

Bibtex

If this work is useful for your research, please consider citing :

@inproceedings{chen2020ssd,
  title={SSD-GAN: Measuring the Realness in the Spatial and Spectral Domains},
  author={Chen, Yuanqi and Li, Ge and Jin, Cece and Liu, Shan and Li, Thomas},
  booktitle={AAAI},
  year={2021}
}

Acknowledgement

The code used in this research is based on mimicry.

Contact

Feel free to contact me if there is any questions (cyq373@pku.edu.cn).