Skip to content

[Usage]: Get Generation Metrics vllm v0.9. #20619

Open
@AI-jitsu

Description

@AI-jitsu

Your current environment

The output of `python collect_env.py`

How would you like to use vllm

I know that the variable Metrics is deprecated in the output of a model generation. But I absolutely need to have them (it is always at None now). Is there any way to be able to get the Metrics ? It involves metrics related to throughput, time to first token, etc... and should be related to each generation request and not a general profiling.

Before submitting a new issue...

  • Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

Metadata

Metadata

Assignees

No one assigned

    Labels

    usageHow to use vllm

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions