Repository files navigation

PyTorch3D provides efficient, reusable components for 3D Computer Vision research withPyTorch.

Key features include:

• Data structure for storing and manipulating triangle meshes
• Efficient operations on triangle meshes (projective transformations, graph convolution, sampling, loss functions)
• A differentiable mesh renderer
• Implicitron, seeits README, a framework for new-view synthesis via implicit representations. (blog post)

PyTorch3D is designed to integrate smoothly with deep learning methods for predicting and manipulating 3D data.For this reason, all operators in PyTorch3D:

• Are implemented using PyTorch tensors
• Can handle minibatches of hetereogenous data
• Can be differentiated
• Can utilize GPUs for acceleration

Within FAIR, PyTorch3D has been used to power research projects such asMesh R-CNN.

See ourblog post to see more demos and learn about PyTorch3D.

For detailed instructions refer toINSTALL.md.

PyTorch3D is released under theBSD License.

Get started with PyTorch3D by trying one of the tutorial notebooks.

Deform a sphere mesh to dolphin	Bundle adjustment

Render textured meshes	Camera position optimization

Render textured pointclouds	Fit a mesh with texture

Render DensePose data	Load & Render ShapeNet data

Fit Textured Volume	Fit A Simple Neural Radiance Field

Fit Textured Volume in Implicitron	Implicitron Config System

Learn more about the API by reading the PyTorch3Ddocumentation.

We also have deep dive notes on several API components:

• Heterogeneous Batching
• Mesh IO
• Differentiable Rendering

We have created a short (~14 min) video tutorial providing an overview of the PyTorch3D codebase including several code examples. Click on the image below to watch the video on YouTube:

We welcome new contributions to PyTorch3D and we will be actively maintaining this library! Please refer toCONTRIBUTING.md for full instructions on how to run the code, tests and linter, and submit your pull requests.

• main branch: actively developed, without any guarantee, Anything can be broken at any time
  • REMARK: this includes nightly builds which are built frommain
  • HINT: the commit history can help locate regressions or changes
• backward-compatibility between releases: no guarantee. Best efforts to communicate breaking changes and facilitate migration of code or data (incl. models).

PyTorch3D is written and maintained by the Facebook AI Research Computer Vision Team.

In alphabetical order:

• Amitav Baruah
• Steve Branson
• Krzysztof Chalupka
• Jiali Duan
• Luya Gao
• Georgia Gkioxari
• Taylor Gordon
• Justin Johnson
• Patrick Labatut
• Christoph Lassner
• Wan-Yen Lo
• David Novotny
• Nikhila Ravi
• Jeremy Reizenstein
• Dave Schnizlein
• Roman Shapovalov
• Olivia Wiles

If you find PyTorch3D useful in your research, please cite our tech report:

@article{ravi2020pytorch3d,author ={Nikhila Ravi and Jeremy Reizenstein and David Novotny and Taylor Gordon                  and Wan-Yen Lo and Justin Johnson and Georgia Gkioxari},title ={Accelerating 3D Deep Learning with PyTorch3D},journal ={arXiv:2007.08501},year ={2020},}

If you are using the pulsar backend for sphere-rendering (thePulsarPointRenderer orpytorch3d.renderer.points.pulsar.Renderer), please cite the tech report:

@article{lassner2020pulsar,author ={Christoph Lassner and Michael Zollh\"ofer},title ={Pulsar: Efficient Sphere-based Neural Rendering},journal ={arXiv:2004.07484},year ={2020},}

Please see below for a timeline of the codebase updates in reverse chronological order. We are sharing updates on the releases as well as research projects which are built with PyTorch3D. The changelogs for the releases are available underReleases, and the builds can be installed usingconda as per the instructions inINSTALL.md.

[Oct 31st 2023]: PyTorch3Dv0.7.5 released.

[May 10th 2023]: PyTorch3Dv0.7.4 released.

[Apr 5th 2023]: PyTorch3Dv0.7.3 released.

[Dec 19th 2022]: PyTorch3Dv0.7.2 released.

[Oct 23rd 2022]: PyTorch3Dv0.7.1 released.

[Aug 10th 2022]: PyTorch3Dv0.7.0 released with Implicitron and MeshRasterizerOpenGL.

[Apr 28th 2022]: PyTorch3Dv0.6.2 released

[Dec 16th 2021]: PyTorch3Dv0.6.1 released

[Oct 6th 2021]: PyTorch3Dv0.6.0 released

[Aug 5th 2021]: PyTorch3Dv0.5.0 released

[Feb 9th 2021]: PyTorch3Dv0.4.0 released with support for implicit functions, volume rendering and areimplementation of NeRF.

[November 2nd 2020]: PyTorch3Dv0.3.0 released, integrating the pulsar backend.

[Aug 28th 2020]: PyTorch3Dv0.2.5 released

[July 17th 2020]: PyTorch3D tech report published on ArXiv:https://arxiv.org/abs/2007.08501

[April 24th 2020]: PyTorch3Dv0.2.0 released

[March 25th 2020]:SynSin codebase released using PyTorch3D:https://github.com/facebookresearch/synsin

[March 8th 2020]: PyTorch3Dv0.1.1 bug fix release

[Jan 23rd 2020]: PyTorch3Dv0.1.0 released.Mesh R-CNN codebase released:https://github.com/facebookresearch/meshrcnn