The code base is no longer maintained.

Please check here for an improved implementation of MUNIT: https://github.com/NVlabs/imaginaire/tree/master/projects/munit

MUNIT: Multimodal UNsupervised Image-to-image Translation

License

Copyright (C) 2018 NVIDIA Corporation. All rights reserved. Licensed under the CC BY-NC-SA 4.0 license (https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode).

For commercial use, please consult NVIDIA Research Inquiries.

Code usage

Please check out the user manual page.

Paper

Xun Huang, Ming-Yu Liu, Serge Belongie, Jan Kautz, "Multimodal Unsupervised Image-to-Image Translation", ECCV 2018

Results Video

Edges to Shoes/handbags Translation

Animal Image Translation

Street Scene Translation

Yosemite Summer to Winter Translation (HD)

Example-guided Image Translation

Other Implementations

MUNIT-Tensorflow by Junho Kim

MUNIT-keras by shaoanlu

Citation

If you find this code useful for your research, please cite our paper:

@inproceedings{huang2018munit,
  title={Multimodal Unsupervised Image-to-image Translation},
  author={Huang, Xun and Liu, Ming-Yu and Belongie, Serge and Kautz, Jan},
  booktitle={ECCV},
  year={2018}
}