- Notifications
You must be signed in to change notification settings - Fork26k
The Ultimate Guide to PyTorch Contributions
Thank you for considering contributing to PyTorch. The contribution process for PyTorch is similar to most open source projects on Github, involving a healthy amount of open discussion between the maintainers, core developers and the broader community.
- All the ways to contribute
- Getting Started with Documentation changes
- Getting Started with Pull Requests
There are many ways to contribute to the PyTorch project, and we value them all. Most open source contributors start with small improvements, and ramp up the scale of their work as they become more familiar with the project and the community. Here are some pathways to becoming a PyTorch contributor, listed in no particular order:
Community Discussions
PyTorch has an active community of users, researchers, and developers. We tend to congregate on thePyTorch User Forum andPyTorch Developer Forum. Participating in conversations here is a great way to learn about PyTorch and contribute your expertise to the community.PyTorch developers also engage in discussions around design and feature-level changes at the PyTorch RFC repo and Slack channel (read more).
Tutorials
If you are already a user of PyTorch, good places to start contributing are the PyTorch tutorials. The PyTorch tutorials are mainly authored by community members, and we always welcome new contributions or updates to existing tutorials. Learn more aboutcontributing to PyTorch tutorials.Documentation
We aim to produce high-quality documentation, but typos or other inaccuracies may creep in. If you find something to fix or improve, please send in a pull request. Thedocs guide contains resources for you to start submitting documentation updates.Reporting Issues
If you happen to run into some unexpected behavior, you can help by creating an issue (if a similar one doesn't already exist on theissue list). Use the Bug Report template and supply as much information as you can, and any additional insights/guesses you might have. When reporting an issue, it's helpful to narrow down the problematic behavior to a minimal reproducible example. SeeFinding or Reporting Issues to get started.Reproducing and Investigating Issues
Another valuable way to contribute is byreproducing open issues. Sometimes the problematic behavior may be isolated to specific environments, or the reported issue is too broad. By repro'ing, you can add details about the problematic behavior, perhaps even further narrow it down towards a root cause, and offer your insights that can help developers to troubleshoot the problem.Proposing New Features
We welcome ideas for new features in PyTorch! A great way to share it with the community is bydrafting an RFC (request for comments), especially if you have fleshed out the design and would like it reviewed. For more casual ideas, theDev-Discuss forum is a good place to start.Submitting Pull Requests
Fixing existing issues or implementing new features require changes to the codebase. We accept new code contributions via Github Pull Requests. We can't accept any changes unless you sign thePyTorch Contributors License Agreement. You will need to set up your developer environment, find an issue to work upon, and submit your changes for review. ThePR guide below contains resources for you to start submitting your changes to PyTorch.Advocate for PyTorch
As a PyTorch user, you are already a valued member of our community. Using PyTorch in your projects, examples, demos, workshops, and blogs goes a long way in raising awareness for PyTorch and the community. Please reach out tomarketing@pytorch.org for any support with sharing your work.PyTorch documentation is generated from docstrings in the python source using Sphinx. This generated HTML is copied to themain folder in thesite branch ofpytorch/docs repo, and is then served via GitHub pages.
You need tobuild and install PyTorch before you can develop on the docs.
Here are some resources for developing PyTorch docs:
Here are some resources that can help you contribute pull requests.
While you can use any tool you want to work on a PR (including AI coding tools, fuzzers, linters), the PR being submitted is your own responsibility and must follow the appropriate coding practices within our codebase. Maintainers will not review the PR until it is following common coding practices inside the codebase (described below) and will deal with repeated, low quality PR, similarly to spam on the repo. When in doubt about anything, please ask on the issue for guidance.
- Install Prerequisites
- Fork, clone, and checkout the PyTorch source
- Install Dependencies
- Build PyTorch from source
- Tips for developing PyTorch
- PyTorch Workflow Git cheatsheet
I would love to contribute to PyTorch!