Tuesday, September 27
Shadow

Announcing the PyTorch Foundation to Accelerate Advances in AI Research

Since 2016, when we partnered with the AI ​​community to create the TorchPy an AI research framework, open collaboration has been critical to its success. With thousands of contributors who have built over 150,000 projects on it, PyTorch has become one of the leading research and production platforms in the AI ​​community.

Today, Mark Zuckerberg announced that the project will move to a new PyTorch Foundation, which will be part of the nonprofit Linux Foundation, a technology consortium whose main mission is the collaborative development of open-source software.

The creation of the PyTorch Foundation ensures that decisions will be made transparently and openly by a diverse group of board members for many years to come. The The governing body will consist of representatives from AMD, Amazon Web Services, Google Cloud, Meta, Microsoft Azure and Nvidia, with the intention of expanding over time.

Our Commitment to Community-Based AI Research

PyTorch was built with an open-source, community-based ethos, and that won’t change with the transition to the Foundation. When researchers and developers open up their code, others around the world can share their work, learn from each other’s progress, and then contribute back to the AI ​​community.

Going forward, framework contributors will benefit from strong governance, diverse leadership, and additional investments provided by new PyTorch Foundation partners. The Foundation will endeavor to respect four principles: to remain open, to maintain a neutral brand image, to remain fair and to forge a strong technical identity. One of his main priorities will be to maintain a clear separation between the business and technical governance of PyTorch.

At Meta, we will continue to invest in PyTorch and use it as the primary framework for our AI research and production. The transition itself will mean no changes to PyTorch’s code, core project, or developer operating models.

We have worked steadily to nurture the community growth that has fueled PyTorch’s success, engaging hundreds of engineers in the framework and supporting product development and community outreach.

Open science is at the heart of our AI work, whether release code for large language models, self-supervised computer vision systems,new innovative datasets, embodied AI platforms and much more. We believe this approach enables the fastest progress in building and deploying new systems that will meet real-world needs and answer fundamental questions about the nature of intelligence. With the creation of the PyTorch Foundation, the entire AI community is able to advance the field in countless exciting new ways.