PyTorch Implementation of Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
imgs init Jan 18, 2018
test init Jan 18, 2018
.gitignore clean up Jan 18, 2018
AdaptiveInstanceNormalization.py init Jan 18, 2018
LICENSE Create LICENSE Jan 18, 2018
README.md fix README. Jan 18, 2018
test_cifar.py init Jan 18, 2018

README.md

Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization in PyTorch

This is an unofficial PyTorch implementation of Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization by Harry Yang. The code is adapted from author's torch implementation AdaIN-style.

The code only contains inference (no training part yet). The decoder model and the vgg model are converted from the torch model using this torch-to-pytorch-convertion, which, together with the original torch model, could be downloaded from here. The code was intended to take two random images from CIFAR10 and use them as the content image and style image respectively. We tested the output such that it is identical with the result of the author's torch implementation.

An example of the result is like this (the content image, the style image and the style transfer result):

cifar10 cifar10 cifar10