Looking at the Pytorch Foundation page [1] and also in the article, I see that the following:<p>> The PyTorch Foundation will boast a wide-ranging governing board composed of representatives from AMD, Amazon Web Services, Google Cloud, Meta, Microsoft Azure, and Nvidia, with the intention to expand further over time.<p>As an academic myself who does research on PyTorch, I wonder if there should have been more academics involved in the guidance or governance of PyTorch, especially given how much the future of machine learning may depend on choices made by such frameworks. Maybe it is unfounded, but I fear
that over-optimizing for large industry uses and forgetting what it is like to run a single network training on your desktop GPU might be detrimental to PyTorch and machine learning as a whole.<p>[1] <a href="https://pytorch.org/foundation" rel="nofollow">https://pytorch.org/foundation</a>
This is a really big announcement. Kudos to Meta. I use PyTorch daily at work and am very happy with both its current state and where it appears to be headed. PyTorch makes deep learning accessible, understandable, and portable, and I love it.