✦
Summary
GNM, the Generative aNthropometric Model and Ecosystem from Google, is an open ecosystem for parametric human models and related perception stacks. The repository starts with GNM Head, a high-fidelity statistical 3D model for human head and face geometry intended for computer vision, graphics, generative AI, and digital human research.
Key Features
Provides an open ecosystem for parametric human models and perception stacks.
Starts with GNM Head, a high-fidelity statistical 3D head and face model.
Offers disentangled control over identity, expression, and head pose.
Models internal anatomy including eyeballs, teeth, and tongue.
Supports NumPy, JAX, PyTorch, and TensorFlow backends.
Includes semantic parameter sampling for controlled human head variation.
Uses a permissive Apache 2.0 license suitable for commercial and non-commercial applications.
Targets computer vision, computer graphics, generative AI, and digital-human workflows.
About GNM
GNM Head provides disentangled control over identity, expression, and head pose, while also modeling internal anatomy such as eyeballs, teeth, and tongue. The package is designed for multi-framework use with NumPy, JAX, PyTorch, and TensorFlow backends, plus semantic parameter sampling for generating and analyzing human head variations.
GNM is useful for researchers and developers building face reconstruction, animation, synthetic data, avatars, digital humans, and 3D perception systems. The repository is released under Apache 2.0 and presents GNM Head as the first step toward a broader suite of statistical human models and analysis technology.