DeOldify
Open-source deep learning that colorizes black-and-white photos and video.
Retired
DeOldify appears to be discontinued or shut down — repo archived Oct 2024; community forks (ReOldify) continue This page is kept as an archive record and is excluded from search engines.
About
DeOldify is an open-source deep learning project created by Jason Antic that colorizes and restores old black-and-white images and film footage. It uses generative adversarial networks with a self-attention U-Net to predict realistic colors, and it became one of the most widely used AI colorization tools after its release.
The original repository was archived in October 2024, but the project lives on through community forks such as ReOldify, which modernizes the stack and adds newer colorization engines. It remains a reference implementation for GAN-based image colorization.
What it can do
- Image colorization with Artistic and Stable models
- Video colorization with temporal consistency
- NoGAN training approach for realistic results
- Open-source weights and code
Setup
Clone the GitHub repository, install the Python dependencies, download the pretrained model weights, then run the notebook or CLI on your images.
Prerequisites
Python 3.7 or newer, PyTorch, and a CUDA-capable GPU for reasonable speed.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 131; Trending score 72/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.