Skip to content

PyTorch implementation of "Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization" by Xun Huang, Serge Belongie.

License

Notifications You must be signed in to change notification settings

aadhithya/AdaIN-pytorch

Repository files navigation

AdaIN-pytorch

PyTorch implementation of "Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization" by Xun Huang, Serge Belongie. arxiv.

0. References

1. Set up and train model

1.1 install requirements

pip install -r requitements.txt

1.2 get the datasets and unzip them to desired location

1.3 Train model

  • to start training: python main.py train path/to/coco path/to/wikiart [OPTIONS]

  • to change training parameters and options: python main.py train --help

2. Style Transfer using pre-trained model

  • to run inference: python main.py infer [OPTIONS]
  • info on supported options: python main.py infer --help

3. Some Results

  • All images were resized to 1024x1024 for inference. The 1024x1024 outputs are interpolated to the original size using bilinear interpolation.
Content Img Style Img Output

About

PyTorch implementation of "Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization" by Xun Huang, Serge Belongie.

Resources

License

Stars

Watchers

Forks

Packages

No packages published