Authors: Yingying Deng, Fan Tang, XingjiaPan, Weiming Dong, Chongyang Ma, Changsheng Xu

This paper is proposed to achieve unbiased image style transfer based on the transformer model. We can promote the stylization effect compared with state-of-the-art methods. This repository is the official implementation of SyTr^2 : Image Style Transfer with Transformers.

• python 3.6
• pytorch 1.4.0
• PIL, numpy, scipy
• tqdm
  

Pretrained models: vgg-model, decoder[Coming SOON], Transformer_module [Coming SOON]
Please download them and put them into the floder ./experiments/

python test.py  --content_dir input/content/ --style_dir input/style/    --output out

Style dataset is WikiArt collected from WIKIART

content dataset is COCO2014

python train.py --style_dir ../../datasets/Images/ --content_dir ../../datasets/train2014 --save_dir models/ --batch_size 8

If you find our work useful in your research, please cite our paper using the following BibTeX entry ~ Thank you ^ . ^. Paper Link pdf

@inproceedings{deng2021stytr2,
      title={StyTr^2: Image Style Transfer with Transformers}, 
      author={Yingying Deng and Fan Tang and Xingjia Pan and Weiming Dong and Chongyang Ma and Changsheng Xu},
      booktitle={CVPR},
      year={2022},
}