D
DDColor
Description
DDColor is the official implementation of an ICCV 2023 paper on photo-realistic image colorization with dual decoders. It adds color to black-and-white images and, per the authors, can also recolor game landscapes into realistic scenes. Code is published under the Apache 2.0 license; weights are on Hugging Face and ModelScope, with more versions listed in the model zoo. Online demos are hosted on ModelScope and Replicate, and a local Gradio demo is included. The page lists Python and PyTorch as requirements but does not state GPU memory. It describes images only: video colorization and the model's limitations are not stated.
Pros and cons
Pros
- Apache 2.0 license, commercial use allowed
- Open weights on Hugging Face and ModelScope
- Online demos and a local Gradio demo
- Several model versions to choose from
Cons
- Photos only, video colorization is not described
- Hardware requirements are not stated by the authors
- Limitations of the model are not described
- Model is from 2023, no newer version is announced on the page
Key facts
- License
- Apache 2.0
- Commercial use
- Allowed
- Developer
- DDColor authors
- Checked on
- Oct 21, 2026
Editorial score
How strongly the AI Expertise editors recommend this open model for each of its tasks right now, on a 0–10 scale. The score combines output quality, capabilities, ease of launch, license openness and freshness.