Skip to content

placeholder for tracking saturated cast logic to float8_e4m3fn #1912

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
vkuzo opened this issue Mar 17, 2025 · 0 comments
Open

placeholder for tracking saturated cast logic to float8_e4m3fn #1912

vkuzo opened this issue Mar 17, 2025 · 0 comments
Assignees
@vkuzo

Comments

@vkuzo
Copy link
Contributor

vkuzo commented Mar 17, 2025 •
edited
Loading

Today the cast to torch.float8_e4m3fn is unsaturated in PyTorch eager mode, and saturated in triton. For now creating a placeholder to improve this, and we might move this to PyTorch core after some discussions.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@vkuzo vkuzo

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@vkuzo