Skip to content

ahmad-PH/pet-colorizer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Pet Colorizer

This application will accept an input grayscale image of a pet (it only can handle dogs and cats for now), and colorize it for you! It was created using the fast.ai library, and most of its code actually comes from Lesson 7 of the fastai course. My main contribution has been to change the model introduced in the course from being an image restorer (turning low resolution images to high resolution ones) to an image colorizer. Also, I wrote a python script that encapsulates the whole AI stuff inside and directly converts an input image file into a colorized output image file.

Here are some examples of what this application can do:





Getting Started

Unfortunately, I haven't had the chance to turn this project into a web application yet, due to Now-Zeit changing their API in a way that breaks all fastai deployments, and being unable to find any other free deployment services to quickly deploy my application. But I will be working on it, and will change the repo to reflect any changes that I'm able to make.

For now, if you want to try colorizing some photos yourself you have to clone this project, change directory to its root directory and then run

python colorizer.py <filename-of-grayscale-image>

the application will then create the colorized image in its root directory (which should be your current directory).

Please Note that the first time you run the "colorizer" script, it will attempt to download the model (as github wouldn't allow for such a huge file to be uploaded) and that might take a while (the download size is approximately 250MB).

Prerequisites

You will need to install the fast.ai library as well as python-wget to run this project. Note that you do not need to use the GPU for running this project, so don't bother setting up the environment for it if you haven't done so already.

Authors

Acknowledgments

  • Thanks to all the amazing people at fast.ai who have made creating such projects a breeze :)

Releases

No releases published

Packages

No packages published