Skip to content

Conversation

@kitaharatomoyo
Copy link
Collaborator

  1. sync in dp group, when all dp group does not contain prefill query, then all dp node forward with cuda_graph;
  2. when pd fusion, moe model prefill with deepep normal and decode with deepep low latency

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