Pytorch implementation of our paper: "Zero-shot unsupervised image-to-image translation via exploiting semantic attributes".
you can install all the dependencies by
pip install -r requirements.txt
- Download CUB dataset.
- Unzip the birds.zip at
./dataset
.
- To view training results and loss plots, run
python -m visdom.server -p 8080
and click the URL http://localhost:8080.
- Run
bash ./scripts/train_bird.sh
- Run
bash ./scripts/test_bird.sh
- The testing results will be saved in
checkpoints/{exp_name}/results
directory.
If this work is useful for your research, please consider citing :
@article{CHEN2022104489,
title = {Zero-shot unsupervised image-to-image translation via exploiting semantic attributes},
journal = {Image and Vision Computing},
pages = {104489},
year = {2022},
issn = {0262-8856},
doi = {https://doi.org/10.1016/j.imavis.2022.104489},
author = {Yuanqi Chen and Xiaoming Yu and Shan Liu and Wei Gao and Ge Li}
}
The code used in this research is inspired by DMIT and FUNIT.
Feel free to contact me if there is any questions ([email protected]).