#6565th of 75 in ML frameworks and librariesFor image, video and design
pytorch3d
PyTorch library for differentiable 3D mesh operations and rendering
- Stars
- 10k
- License
- custom license
- Last commit
- Oct 2026
- Last release
- Nov 2025
- Language
- Python
9,782 on 31 Mar 20269,986 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
PyTorch3D provides triangle mesh data structures, mesh operations (projective transforms, graph convolution, sampling, losses) and a differentiable mesh renderer, all built on PyTorch tensors with heterogeneous minibatch support. It also includes Implicitron, a framework for novel-view synthesis with implicit representations. It targets 3D computer vision research, with tutorial notebooks for mesh fitting, NeRF and bundle adjustment.
Who it is for: 3D computer vision researchers working in PyTorch
Strengths
- All operators are differentiable, batch heterogeneous data and can run on GPU
- Differentiable mesh renderer plus point cloud and volume rendering
- Tutorial notebooks cover NeRF, bundle adjustment, textured mesh fitting and ShapeNet loading
- Includes Implicitron framework for implicit-representation view synthesis
Weaknesses
- No backward-compatibility guarantee between releases
- main branch and nightly builds may break at any time
- README news list stops at v0.7.5 (Oct 2023)
- License detected as NOASSERTION; README says BSD
What it needs
- GPU optional
- Needs PyTorch
- Models: triangle meshes, point clouds, volumes, implicit representations (NeRF)
Also in ML frameworks and libraries
See all 75| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 92 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 91 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 90 out of 100 |