https://github.com/wesleyaldrich/couleurs
Web application RESTful API for image colorization AI. The application is supposed to receive inputs in form of images, then return the colorized same image.
https://github.com/wesleyaldrich/couleurs
ai colorizer deoldify flask
Last synced: 2 months ago
JSON representation
Web application RESTful API for image colorization AI. The application is supposed to receive inputs in form of images, then return the colorized same image.
- Host: GitHub
- URL: https://github.com/wesleyaldrich/couleurs
- Owner: wesleyaldrich
- License: mit
- Created: 2024-08-28T07:51:48.000Z (almost 2 years ago)
- Default Branch: stable
- Last Pushed: 2024-10-01T07:28:53.000Z (over 1 year ago)
- Last Synced: 2025-06-04T00:56:37.327Z (about 1 year ago)
- Topics: ai, colorizer, deoldify, flask
- Language: Python
- Homepage:
- Size: 8.97 MB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Couleurs
**Couleurs** is a simple image colorizing project. The application is supposed to receive inputs in form of images, then return the colorized same image.
This project utilizes the model of **DeOldify by Jason Antic** as the AI to colorize images.
## Team Members
- **Rachell Vannessa Christian** as Project Coordinator
- **Kalista Gabriela Willies** as Technical and Research Analyst
- **Monica Agustina Chandra** as UI and UX Designer
- **Michael Onasis Hasri** as Major Front-end Developer
- **Wesley Aldrich** as Back-end Developer, Minor Front-end Developer
## Website
In this repository, I'm creating the web application using **Flask**. The goal of this project is:
- To make sure I completely understand how **DeOldify AI** works.
- To simply learn how to use **Flask** framework in building simple web applications.
## How to use
Clone this repository into your device, then run the **"app.py"** file from your terminal.
If any error pops up, it might be because some required libraries are not in your device.
Install all the dependencies that are "_unrecognized_".
### Cloning this repository
``` bash
git clone https://github.com/wesleyaldrich/Couleurs.git
cd Couleurs
```
"Couleurs" is the web application root directory, which is now cloned into your device.
### Installing DeOldify Model
- Download an AI Model Weights for our image colorizer.
- In this project, we will use the Artistic Completed Generator Weights from DeOldify.
[Download Here](https://data.deepai.org/deoldify/ColorizeArtistic_gen.pth)
- After it is downloaded, the file should be named `ColorizeArtistic_gen.pth`. Replace the existing dummy file in `Couleurs/ai/models/`!
- Thus, the file path for the model weights should be `Couleurs/ai/models/ColorizeArtistic_gen.pth`.
### Installing dependencies
If your Windows device is compatible with Python 3.7.6, you can install the required libraries perfectly by running:
``` bash
pip install -r requirements_freeze.txt
```
Otherwise, you can run this:
``` bash
pip install -r requirements.txt
```
Or manually by:
``` bash
pip install flask matplotlib pandas scipy requests fastprogress torch torchvision opencv-python ffmpeg yt_dlp ipython
```
If you're using Linux/Ubuntu, you might also need:
``` bash
sudo apt install libgl1-mesa-glx
```
- Note: Consider using virtual environment as well.
### Run the web application
- For Linux/Ubuntu:
```
python3 app.py
```
- For Windows:
```
python app.py
```
- Or you can adjust with your own operating system or preference.
By default, you should see Couleurs local web app running in `http://127.0.0.1:8080/`.
---
- DeOldify by Jason Antic
https://github.com/jantic/DeOldify/