Skip to content

Fix a bug in cuda graph runner#1094

Merged
merrymercy merged 1 commit intomainfrom
fix-cuda-graph
Aug 14, 2024
Merged

Fix a bug in cuda graph runner#1094
merrymercy merged 1 commit intomainfrom
fix-cuda-graph

Conversation

@merrymercy
Copy link
Copy Markdown
Contributor

@merrymercy merrymercy commented Aug 14, 2024

This pr fixes the padding on the sequence length for cuda graph. Otherwise, it sometimes reports illegal memory access.

@merrymercy merrymercy marked this pull request as ready for review August 14, 2024 10:10
@merrymercy merrymercy requested review from hnyls2002 and zhyncs August 14, 2024 10:11
@merrymercy merrymercy merged commit 8f790ac into main Aug 14, 2024
@merrymercy merrymercy deleted the fix-cuda-graph branch August 14, 2024 10:25
timethink pushed a commit to timethink/sglang that referenced this pull request Mar 9, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant