Spaces:
Running
on
Zero
Running
on
Zero
Allow cuda tf32 matmul to optimize
#39
by
billloic
- opened
No description provided.
Thank you for the pull request! I have added torch.backends.cuda.matmul.allow_tf32 = True to the code in the Space. This change is expected to improve performance in the hardware environment provided by Hugging Face. Thank you for your input!