Skip to content

Conversation

dagrayvid
Copy link
Collaborator

This PR fixes two small bugs.

  • Double counting of prompt tokens in the calculation of total tokens_per_second
  • The warmup percent was not working when using max duration (worked for max requests). Traced this to a - which should have been a + in benchmark/aggregator.py

@dagrayvid dagrayvid requested a review from markurtz May 27, 2025 15:19
@markurtz markurtz merged commit 66017f5 into vllm-project:main May 29, 2025
9 of 10 checks passed
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.

2 participants