#157157th of 206 in ML frameworks and librariesFor image, video and design, 3d
kaolin
PyTorch library for 3D deep learning: meshes, splats, rendering, physics
- Stars
- 5.2k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Aug 2025
- Language
- Python
5,046 on 21 Apr 20265,195 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Kaolin is NVIDIA's PyTorch library of GPU-optimized modules for 3D deep learning research. It covers differentiable rendering (DIB-R, nvdiffrast, easy_render PBR), 3D Gaussian splat containers with PLY/USD I/O, Structured Point Clouds (GPU octrees), representation conversions, quaternion math, and Warp-based Simplicits physics for meshes, splats and point clouds.
Who it is for: 3D vision and graphics researchers working in PyTorch
Strengths
- Covers rendering, conversions, physics, and USD I/O in one PyTorch API
- Many tutorial notebooks plus API docs on readthedocs
- Apache-2.0 including FlexiCubes, now relicensed from non-commercial
- Prebuilt pip wheels for specific torch/CUDA combinations
Weaknesses
- Wheels are published per torch/CUDA pair; other combinations need a source build
- Web client-server framework only on a prerelease branch, not merged to master
- Newton coupling is marked experimental
- Last release 0.18.0 is from 2025-08; recent features exist only on master
What it needs
- GPU required
- Needs PyTorch, CUDA, NVIDIA Warp
- Models: PLY, USD, OBJ, GLTF
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in ML frameworks and libraries
See all 206| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | ultralyticsPython library and CLI for training and running YOLO vision models | 94 out of 100 |
| 3 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |