# NVIDIAGameWorks/kaolin

A PyTorch Library for Accelerating 3D Deep Learning Research

Repository: https://github.com/NVIDIAGameWorks/kaolin
Canonical: https://ross.abutalabs.com/products/kaolin
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pytorch, 3d-deep-learning, neural-networks, differentiable-rendering, artificial-intelligence, camera-api, cuda, rasterization, differentiable-lighting, gaussian-splatting, interactive-visualizations, nvidia-warp, physics-simulation
Last push: 2026-08-24T21:35:41+00:00

## Health v2 (maintenance only)
Score: 68/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 10, longevity 100
- inputs: {"age_days": 2484, "days_push": 9, "days_rel": 391, "gap_med": 260, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5161, forks 628 (observed 2026-08-28T04:09:11.484581+00:00)

## What it is
Kaolin is NVIDIA's PyTorch library of GPU-optimized modules for 3D deep learning research, covering meshes, point clouds, and 3D Gaussian splats. It provides differentiable rendering, representation conversions, physics simulation, octree acceleration, and I/O/visualization tools.

## Use cases
- train neural networks on 3d meshes in pytorch
- differentiable rendering for 3d reconstruction
- work with 3d gaussian splatting models
- convert between 3d representations like meshes point clouds and splats
- simulate physics on 3d scenes with deep learning
- load and export 3d assets in usd or ply format
- accelerate 3d deep learning research on gpu

## When to choose
- you need GPU-accelerated 3D deep learning building blocks within PyTorch
- you want differentiable rendering or Gaussian splatting support
- you need conversions and I/O across multiple 3D representations

## When to avoid
- you need a general-purpose 3D modeling or game engine rather than a research library
- your project does not use PyTorch or CUDA-capable GPUs
- you only need simple non-differentiable 3D rendering

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, graphics, simulation, image-processing, data-visualization
- domain: deep-learning, computer-vision, graphics, artificial-intelligence, simulation
- platform: python, windows, cross-platform
- tags: pytorch, 3d-deep-learning, differentiable-rendering, gaussian-splatting, mesh-processing, point-clouds, cuda, nvidia, usd, physics-simulation, gpu, linux

## Member repositories
- NVIDIAGameWorks/kaolin (main) score 68

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:11.484581+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:01:53.869027+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIAGameWorks/kaolin (fetched 2026-08-28T04:09:11.484581+00:00, sha 34bcd565887e)
- Data as of 2026-08-30T08:39:29.467469+00:00.