Skip to content
SPECIFICTION
Get listed

DDColor

github.com

Free plan LinuxAPISelf-hosted
Visit github.com
Spec score 7.9 #13 of 26 in AI Photo Colorization Software

What it is

DDColor is a free model for adding color to historical black-and-white photos and recoloring anime-game landscapes. It uses multi-scale visual features to optimize learnable color tokens for automatic colorization. The repository offers paper, ModelScope, artistic and tiny pretrained variants; the maker recommends ModelScope for images outside ImageNet and identifies tiny as the lightest option. Local inference scripts and a runnable Gradio demo are documented, alongside inference guidance for Hugging Face Hub and ModelScope. The repository also links to Replicate demos and an API. Users can train with ImageNet or a custom dataset, and the documentation covers ONNX export with an example using the tiny model. Requirements are Python 3.7 or later and PyTorch 1.7 or later. The repository includes an Apache License, Version 2.0, and lists no paid plans or product usage limits. One stated limitation is that colorfulness loss can result in unattractive color blocks, including red artifacts.

Spec sheet

Purpose
DDColor colorizes historical black-and-white photos and can recolor anime-game landscapes.Read 4 Oct 2026 · github.com
Integrations
Inference instructions cover Hugging Face Hub and ModelScope, and the repository links to Replicate demos and an API.Read 4 Oct 2026 · github.com
Security
The repository page does not state security controls or compliance certifications.Read 4 Oct 2026 · github.com
Support
The repository directs users to online demos and documentation but does not state a dedicated support service.Read 4 Oct 2026 · github.com
License
The repository includes an Apache License, Version 2.0.Read 4 Oct 2026 · github.com
Export
The repository documents ONNX export, with an example using the tiny pretrained model.Read 4 Oct 2026 · github.com
Local use
The README documents local inference scripts and a runnable Gradio demo.Read 4 Oct 2026 · github.com
Maker attribution
The paper authors are listed as affiliated with DAMO Academy, Alibaba Group.Read 4 Oct 2026 · github.com
Method
The model uses multi-scale visual features to optimize learnable color tokens for automatic image colorization.Read 4 Oct 2026 · github.com
Model guidance
The maker recommends the ModelScope model by default for images outside ImageNet and describes the tiny model as the most lightweight version.Read 4 Oct 2026 · github.com
Model options
The repository provides paper, ModelScope, artistic, and tiny pretrained model variants.Read 4 Oct 2026 · github.com
Notable limitation
The maker cautions that colorfulness loss can produce unpleasant color blocks, including red artifacts.Read 4 Oct 2026 · github.com
Requirements
The listed requirements are Python 3.7 or later and PyTorch 1.7 or later.Read 4 Oct 2026 · github.com
Training
Users can train the model using ImageNet or a custom dataset, with training dependencies and pretrained ConvNeXt and InceptionV3 weights.Read 4 Oct 2026 · github.com

Plans and prices

DDColor open-source software Free No paid plans or product usage limits stated Read 4 Oct 2026 · github.com

Every DDColor plan and price →

Strong points and limits

Strong points

  • Free under Apache License, Version 2.0
  • Includes four pretrained model variants
  • Documents local inference and a Gradio demo
  • Supports custom-dataset training and ONNX export
  • API access and batch processing are listed

Limits

  • Colorfulness loss can create red artifacts
  • Requires Python 3.7 or later and PyTorch 1.7 or later
  • No dedicated support service is stated

Who it suits

It suits developers and users looking to run or train an image-colorization model locally, through a demo, or via documented inference integrations. The tiny variant is identified as the lightest option, while ModelScope is recommended for images outside ImageNet.

Verdict

DDColor offers multiple pretrained variants and documented paths for inference, training and ONNX export. Be aware of the stated risk of color blocks, including red artifacts, during colorization.

Where Specifiction places it

Alternatives to DDColorAll alternatives →