Manga Colorizer (Standalone GUI Fork)

This project provides a standalone desktop application with a graphical user interface (GUI) for the Manga-Colorizer. It functions as a UI wrapper, removing the original client/server architecture that required running a backend server and a browser-based front end (like Firefox, Safari, or Chrome).

Unlike the original version which colorized images directly on websites, this application allows you to process local image files from any folder on your computer. This provides a streamlined, offline-first workflow for colorizing, denoising, and upscaling your manga pages, allowing you to process entire folders of images with just a few clicks.

TL;DR

• Original: Runs a server + uses your web browser as a plugin to process pages.
• This fork: Runs as a desktop app — no browser, no separate server.
• Original: Processes web pages individually.
• This fork: Can process entire folders at once.

---

📦 Full Installation Guide

Follow these steps to set up and run the application on your local machine.

1️⃣ Clone the Repository

Clone this repository with Git:

git clone https://github.com/sepTN/Manga-Colorizer-GUI
cd Manga-Colorizer-GUI

Or download as a ZIP and extract it.

---

2️⃣ Set Up the Conda Environment

We recommend Conda to manage dependencies.

Create a new environment:

conda create --name manga-colorizer python=3.9

Activate it:

conda activate manga-colorizer

---

3️⃣ Install Required Libraries

Standard Packages:

pip install Flask Flask_Cors matplotlib numpy opencv_python_headless scikit_image einops

PyTorch Installation: choose the correct one for your system.

• Mac (Apple Silicon M1/M2/M4):

  pip install torch torchvision

• Windows/Linux with NVIDIA GPU (CUDA):

  pip install torch torchvision --index-url https://download.pytorch.org/whl/cu118

• Windows/Linux without a dedicated GPU:

  pip install torch torchvision --index-url https://download.pytorch.org/whl/cpu

---

4️⃣ Download the AI Models

Download and place this generator.zip file in Backend/networks/.

• Colorizer Model:
  Generator weights → generator.zip

• Upscaler Models:

  • ESRGAN (default, already included): RealESRGAN_x4plus_anime_6B.pt
  • GigaGAN (optional): I haven’t tested this model myself, so you’re on your own here — in the app’s Advanced Settings, you can set the path to the model file manually.

---

5️⃣ Run the Application

From the main project folder:

python app.py

The GUI will appear—start colorizing your manga! 🎨

---

🙌 Credits

Built on the work of amazing developers:

• BinitDOX – Original Manga-Colorizer
• qweasdd – manga-colorization-v2
• xiaogdgenuine – Manga-Colorization-FJ
• xinntao – Real-ESRGAN
• vatavian – Improved fork with on-the-fly colorization
• iG8R – Testing and feedback
• Forked and streamlined by sepTN for a simpler all-in-one experience.