This extension for AUTOMATIC1111's Stable Diffusion WebUI enables the colorization of old photos and videos. It targets users of Stable Diffusion who wish to restore and enhance historical media, offering a straightforward integration for improved image and video quality.

How It Works

The extension leverages the DeOldify model, a deep learning approach known for its effectiveness in colorizing grayscale images. It integrates directly into the Stable Diffusion WebUI's "Extra" tab, allowing users to upload media and apply colorization with adjustable parameters like "Render Factor."

Quick Start & Requirements

• Install via the "Extensions" tab in the Stable Diffusion WebUI by providing the URL: https://github.com/SpenserCai/sd-webui-deoldify.
• Video colorization requires ffmpeg to be installed (sudo apt install ffmpeg).
• Compatibility is confirmed with Pytorch versions <=1.13.1+cu117 and 2.0.1+cu118, and ROCm 5.5.
• Official documentation and demo links are not explicitly provided in the README.

Highlighted Details

• Supports both photo and video colorization.
• Offers a Discord bot for remote colorization tasks.
• Aims to improve video processing speed.
• Future plans include support for repair options and advanced generation scripting.

Maintenance & Community

The project has seen recent updates, including the release of a self-deployable DeOldifyBot. Community contributions are encouraged via Pull Requests for testing and compatibility reporting. Links to community channels like Discord or Slack are not provided.

Licensing & Compatibility

The README does not explicitly state a license. Given its reliance on Stable Diffusion and DeOldify, users should verify licensing compatibility for commercial use or integration into closed-source projects.

Limitations & Caveats

The extension is still under active development, with features like "repair options" and advanced scripting still pending. Users may need to ensure specific Pytorch and CUDA/ROCm versions for optimal compatibility.