Skip to content

Latest commit

 

History

History
88 lines (65 loc) · 3.44 KB

README.md

File metadata and controls

88 lines (65 loc) · 3.44 KB

Paints Chainer

Paints Chainer is line drawing colorizer using chainer. Using CNN, you can colorize your scketch automatically / semi-automatically .

image

Requirement

If not specified, versions are assumed to be recent LTS version.

  • A Nvidia graphic card supporting CUDA (6.5, 7.0, 7.5, 8.0)
  • Linux: gcc/ g++ 4.8
  • Windows: "Microsoft Visual C++ Build Tools 2015" (NOT "Microsoft Visual Studio Community 2015")
  • Python 3 (3.5 recommended) (Python 2.7 needs modifying web host (at least) )
  • Numpy
  • openCV "cv2" (Python 3 support possible, see installation guide)
  • Chainer
  • CUDA / cuDNN (If you use GPU)

Installation Guide

Option: Docker user

If you have docker, you may check https://hub.docker.com/r/liamjones/paintschainer-docker/ (not supported officially but thanks for volunteering)

Option: Fresh install

If not specified, follow instruction from their official website.

If you failed to perform the following steps, you will see this message. Uninstall chainer and install it again.

 Running command python setup.py egg_info
    Options: {'profile': False, 'annotate': False, 'linetrace': False, 'no_cuda': False}
    Include directories: ['C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v8.0\\include']
    Library directories: ['C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v8.0\\bin', 'C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v8.0\\lib\\x64']
    Microsoft Visual C++ 14.0 is required. Get it with "Microsoft Visual C++ Build Tools": http://landinghub.visualstudio.com/visual-cpp-build-tools

Starting web host

UI is html based. using wPaint.js Server side is very basic python server

boot local server python server.py

access to localhost localhost:8000/

Learning

main code of colorization is in cgi-bin/paint_x2_unet

to train 1st layer using GPU 0 python train_128.py -g 0 to train 2nd layer using GPU 0 python train_x2.py -g 0

DEMO

http://paintschainer.preferred.tech/

License

Source code : MIT License

Pre-trained Model : All Rights Reserved

Pre-Trained Models

Download following model files to cgi-bin/paint_x2_unet/models/

http://paintschainer.preferred.tech/downloads/

(Copyright 2017 Taizan Yonetsuji All Rights Reserved.)

Acknowledgements

This project is powered by Preferred Networks.

Thanks a lot for rezoolab, mattya, okuta, ofk . This project could not be achived without their great support.

Line drawing of top image is by ioiori18.