NVlabs/eg3d

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

3338 stars · 362 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Official PyTorch implementation of EG3D, an efficient geometry-aware 3D generative adversarial network from NVIDIA Research (CVPR 2022). It synthesizes high-resolution multi-view-consistent images and 3D geometry from single-view 2D photo collections using a hybrid explicit-implicit network with neural rendering.

Use cases

• generate multi-view-consistent images of faces or cats
• generate 3D shapes from 2D photographs
• render videos from a pre-trained 3D GAN model
• research 3D-aware image synthesis
• experiment with geometry-aware GAN architectures
• produce novel views of generated objects

When to choose

• you need state-of-the-art 3D-aware GAN synthesis with geometry output
• you have high-end NVIDIA GPUs and want a research-grade reference implementation
• you want to build on StyleGAN2-style generators for 3D content

When to avoid

• you lack CUDA-capable NVIDIA GPUs
• you need a production-ready 3D asset pipeline rather than research code
• you need a permissively licensed library for commercial use without checking NVIDIA licensing terms

Facets

library · maturity maintenance

deep-learning machine-learning image-processing graphics deep-learning computer-vision graphics artificial-intelligence python 3d-gan neural-rendering triplegan stylegan2 3d-aware-synthesis cvpr-2022 pytorch research-code linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="NVlabs/eg3d")