Skip to content

problems with sglang #68

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.

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

Open
eliecerecology opened this issue Nov 25, 2024 · 1 comment
Open

problems with sglang #68

eliecerecology opened this issue Nov 25, 2024 · 1 comment

Comments

@eliecerecology
Copy link

eliecerecology commented Nov 25, 2024

when I run guidellm with the default parameters:
'''
guidellm --target "http://localhost:30020/v1/" --model "meta-llama/Llama-3.1-8B-In
struct" --data-type emulated --data "prompt_tokens=112,generated_tokens=58"
'''

I got warnings that data is not available:

| percentiles | WARNING - No data points available to calculate percentiles.
mean | WARNING - No data points available to calculate mean.

and finally, and Error:
FATAL: exception not rethrown
Aborted (core dumped)

The report is missing the metrics "Time to first token", "Inter token latency", "output token throughput"

any suggestion? I know the warnings are coming from the src/guidelln/core/ folder directory, specifically from the module distribution.py. I do not know why error is happening.

Notice there is no problem with vLLM

@markurtz
Copy link
Member

@eliecerecology we haven't explicitly tested or added support for SGLang. There are a few fixes landing with #91, though, that may fix issues here.

Could you supply an example command for how you're launching SGLang?

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

No branches or pull requests

2 participants