Today we are more than thrilled to welcome PyTorch to the Linux Foundation. Honestly, it’s hard to capture how big a deal this is for us in a single post but I’ll try.
TL;DR — PyTorch is one of the most important and successful machine learning software projects in the world today. We are excited to work with the project maintainers, contributors and community to transition PyTorch to a neutral home where it can continue to enjoy strong growth and rapid innovation. We are grateful to the team at Meta, where PyTorch was incubated and grew into a massive ecosystem, for trusting the Linux Foundation with this crucial effort. The journey will be epic.
Artificial Intelligence, Machine Learning, and Deep Learning are critical to present and future technology innovation. Growth around AI and ML communities and the code they generate has been nothing short of extraordinary. AI/ML is also a truly “open source-first” ecosystem. The majority of popular AI and ML tools and frameworks are open source. The community clearly values transparency and the ethos of open source. Open source communities are playing and will play a leading role in development of the tools and solutions that make AI and ML possible — and make it better over time.
For all of the above reasons, the Linux Foundation understands that fostering open source in AI and ML is a key priority. The Linux Foundation already hosts and works with many projects that are either contributing directly to foundational AI/ML projects (LF AI & Data) or contributing to their use cases and integrating with their platforms. (e.g., LF Networking, AGL, Delta Lake, RISC-V, CNCF, Hyperledger).
PyTorch extends and builds on these efforts. Obviously, PyTorch is one of the most important foundational platforms for development, testing and deployment of AI/ML and Deep Learning applications. If you need to build something in AI, if you need a library or a module, chances are there is something in PyTorch for that. If you peel back the cover of any AI application, there is a strong chance PyTorch is involved in some way. From improving the accuracy of disease diagnosis and heart attacks, to machine learning frameworks for self-driving cars, to image quality assessment tools for astronomers, PyTorch is there.
Originally incubated by Meta’s AI team, PyTorch has grown to include a massive community of contributors and users under their community-focused stewardship. The genius of PyTorch (and a credit to its maintainers) is that it is truly a foundational platform for so much AI/ML today, a real Swiss Army Knife. Just as developers built so much of the technology we know today atop Linux, the AI/ML community is building atop PyTorch – further enabling emerging technologies and evolving user needs. From August 2021 through August 2022, PyTorch counted over 65,000 commits. Over 2,400 contributors participated in the effort, filing issues or PRs or writing documentation. These numbers place PyTorch among the most successful open source projects in history.
Projects like PyTorch that have the potential to become a foundational platform for critical technology benefit from a neutral home. Neutrality and true community ownership are what has enabled Linux and Kubernetes to defy expectations by continuing to accelerate and grow faster even as they become more mature. Users, maintainers and the community begin to see them as part of a commons that they can rely on and trust, in perpetuity. By creating a neutral home, the PyTorch Foundation, we are collectively locking in a future of transparency, communal governance, and unprecedented scale for all.
As part of the Linux Foundation, PyTorch and its community will benefit from our many programs and support communities like training and certification programs (we already have one in the works), to community research (like our Project Journey Reports) and, of course, community events. Working inside and alongside the Linux Foundation, the PyTorch community also has access to our LFX collaboration portal, enabling mentorships and helping the PyTorch community identify future leaders, find potential hires, and observe shared community dynamics.
PyTorch has gotten to its current state through sound maintainership and open source community management. We’re not going to change any of the good things about PyTorch. In fact, we can’t wait to learn from Meta and the PyTorch community to improve the experiences and outcomes of other projects in the Foundation. For those wanting more insight about our plans for the PyTorch Foundation, I invite you to join Soumith Chintala (co-creator of PyTorch) and Dr. Ibrahim Haddad (Executive Director of the PyTorch Foundation) for a live discussion on Thursday entitled, PyTorch: A Foundation for Open Source AI/ML.
We are grateful for Meta’s trust in “passing us the torch” (pun intended). Together with the community, we can build something (even more) insanely great and add to the global heritage of invaluable technology that underpins the present and the future of our lives. Welcome, PyTorch! We can’t wait to get started!
