Skip to content

shaotao-art/style-transfer-pytorch

Repository files navigation

Style Transfer Pytorch

implementation of the paper: arbitrary style transfer in real-time with adaptive instance normalization

Sample Results

different patch size

small_patch.png

large_patch.png

different style weight

different_style_w.png

More samples

sample1.png

sample2.png

sample3.png

sample4.png

Data

prepare content images and style images, you can download style image from here, for content image i use VOC12. you can use any content image you want.

Training

use run.sh to run, you can change config in config file configs/config.py

you can find pretrained model with different style wright in here, in the link, the tensorboard log are uploaded too, you can check it locally.

Testing

use play.ipynb to use pretrained model to do style transfer

TODO

  • i do not add reflective pad in vgg model, so there will be artifacts in borders

Reference

/~https://github.com/naoto0804/pytorch-AdaIN

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published