Out-of-box XLA on Huggingface PyTorch models #86
kevint324
started this conversation in
Project Discussions
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'm not sure if this is the right place to ask since it involves multiple interest groups such as openxla community, pytorch/xla and huggingface.
Unlike tensorflow's effort to push XLA with HF, AFAIK there is little work on PyTorch/XLA and HF integration. And there is virtually zero model could work on XLA out-of-box, espacially with the latest LLMs with the recusive pattern usually trigger endless re-compilation.
Is out-of-box experience a goal or prioritized task of the openxla project?
Thanks
Beta Was this translation helpful? Give feedback.
All reactions