Nikita Shulga is seasoned software engineer and Linux enthusiast. He currently works on PyTorch at Meta, where among other things he is responsible for open-source r
eleases and continuous integration.
How to help yourself and ML community make PyTorch better.
PyTorch is a popular platform for deep learning, favored for its flexibility and intuitive design. However, the true power of PyTorch is not just in its technical prowess but in the vibrant community that surrounds it. This talk aims to explore actionable strategies for individuals and communities to contribute to the PyTorch ecosystem.
Attendees will learn what to do if PyTorch does not work the way they expect it to, delving into the best practices for reporting, debugging, and fixing issues as well as to suggest new functionality. We will discuss how to engage with the community through forums, contribute to open-source projects, and improve documentation. The session will also cover how PyTorch is build and tested and walks attendees thru the life-cycle of an issue or a pull request.