derqwertzus/gnm-head-add-on-for-blender — explained in plain English
Analysis updated 2026-05-18
Generate a base human head for a 3D character by blending gender and ethnicity sliders.
Apply one of twenty facial expressions with adjustable intensity to a head.
Randomize only the eyes or mouth of a head using a painted region mask.
Bake a finished head into an independent object before generating new variations.
| derqwertzus/gnm-head-add-on-for-blender | 0petru/sentimo | agi-eval-official/forte | |
|---|---|---|---|
| Stars | 17 | 17 | 17 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | moderate | hard |
| Complexity | 3/5 | 3/5 | 4/5 |
| Audience | designer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Blender 4.2+, a system Python 3.10-3.12, and an internet connection for a one-time ~170MB environment download.
This is a Blender add-on that lets you generate and sculpt realistic human heads using Google's GNM, or Generative Anthropometric Model, a system trained to produce plausible human head shapes. Instead of manually sculpting a face from scratch, you use sliders to blend traits like gender and ethnicity, choose from twenty facial expressions with adjustable intensity, and watch the head update live in the 3D viewport. A useful feature is region-masked randomization. Six preset masks cover areas like the eyes, nose, mouth, jaw, ears, and back of the head, and when one is active, randomizing the identity only changes that masked area while dimming the rest of the model so you can see exactly what is being edited. You can also paint and save your own custom masks. The generated heads use clean quad-based topology rather than a messy triangulated mesh, and you can toggle visibility of parts like eyes, teeth, tongue, and cornea shells, with hidden parts fully removed rather than just hidden. Once you land on a head you like, the Bake Identity feature copies it into an independent object so you can keep generating new variations without losing your saved ones. The add-on works by installing and running a small local server that handles the actual model computation using NumPy rather than TensorFlow, keeping the whole setup around 170 megabytes. To use it, you need Blender 4.2 or newer and a system Python between versions 3.10 and 3.12 available on your computer, plus an internet connection for the one-time setup where the add-on downloads the GNM model and its dependencies. Installation is done by downloading the extension file from the Releases page and installing it through Blender's extension manager, then clicking Install Environment inside the add-on's preferences.
A Blender add-on that generates and sculpts realistic human heads using Google's GNM model, with sliders for identity, expression, and region-masked edits.
Mainly Python. The stack also includes Python, Blender, NumPy.
The add-on itself is under GPL, which requires derivative works to also share their source code, the underlying GNM model is separately licensed under the more permissive Apache-2.0.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly designer.
This repo across BitVibe Labs
Verify against the repo before relying on details.