sxyu/svox2

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2889 stars · 354 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official PyTorch/CUDA implementation of Plenoxels, a method for optimizing sparse voxel radiance fields for novel view synthesis without neural networks. It supports NeRF-Blender, LLFF, NSVF, and CO3D dataset formats and renders photorealistic 3D scenes from posed images.

Use cases

• reconstruct 3d scenes from photos
• novel view synthesis without neural networks
• train radiance fields faster than nerf
• render free viewpoint videos from images
• generate radiance field datasets for research

When to choose

• you want NeRF-quality novel view synthesis with faster optimization and no MLP
• you have a Linux machine with a CUDA GPU and posed multi-view images
• you need a research baseline or building block for radiance field experiments

When to avoid

• you are on Windows, which is not officially supported
• you need real-time rendering or instant-ngp-style hash-grid speed
• you want a maintained production library rather than a research codebase

Facets

library · maturity maintenance

machine-learning graphics image-processing gpu-computing computer-vision graphics machine-learning python radiance-fields nerf voxel-grids 3d-reconstruction novel-view-synthesis cuda-extension cvpr research linux gpu cuda

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sxyu/svox2")