Image-to-Image Translation with Conditional Adversarial Networks (Pix2pix) implementation in keras
Clone or download
Latest commit a2aea2f Mar 27, 2017
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
pix2pix fixed filter big Mar 27, 2017
.gitignore refractored networks Mar 26, 2017
LICENSE Initial commit Feb 23, 2017
README.md changed blog link Mar 27, 2017
requirements.txt added datagens Mar 26, 2017

README.md

pix2pix-keras

Pix2pix implementation in keras.

Original paper: Image-to-Image Translation with Conditional Adversarial Networks (pix2pix)
Paper Authors and Researchers: Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, Alexei A. Efros

To run

# clone repo and step into root dir
git clone https://github.com/williamFalcon/pix2pix-keras.git
cd pix2pix-keras/pix2pix

# download facades dataset   
python utils/facades_dataset.py

# setup conda(or virtual env)
conda create -n pix python=3.4  
source activate pix

# install requirements
pip install requirements.txt

# start training
python main.py 

About this implementation

A full write-up about this implementation is available at my blog https://www.williamfalcon.com/deeplearningimplementations/pix2pix-keras