/Photorealistic-Style-Transfer

High-Resolution Network for Photorealistic Style Transfer

Primary LanguageJupyter NotebookMIT LicenseMIT

High-Resolution Network for Photorealistic Style Transfer

Code and data for paper "High-Resolution Network for Photorealistic Style Transfer".

This code is based on PyTorch 1.0.0 and Cuda 9.0.

Examples

Here are some results from our algorithm (from left to right are input, style and our output):





Acknowledgement

Our work is inspired by Deep High-Resolution Representation Learning for Human Pose Estimation.

The transfer code is based on Udacity.

Contact

Feel free to contact me if there is any question (Ming Li limingcv@gmail.com).