This project is a comprehensive comparison of the following works:
-
Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization (Huang et Belongie, 2017)
-
Image Style Transfer Using Convolutional Neural Networks (Gatys et al, 2016)
The .py files contain the code with the necessary methods to do style transfer
The folder saved_models/contains the necessary models to be loaded
The demo on how the code can be used to perform style transfer is in Experiments.ipynb, AdaIN.ipynb and Gatys.iypnb
A few content and style samples can be found in content/ and style/ folders
The datasets used can be downloaded through the links available in Chinese_art_dataset and COCO_test2017_2k
-
The code for Gatys et al. approach was inspired by the authors' repo: https://github.com/leongatys/PytorchNeuralStyleTransfer
-
The Code for AdaIN was inspired by the following repo: https://github.com/ziwei-jiang/AdaIN-Style-Transfer-PyTorch/tree/master