Skip to content

Conversation

@nirda7
Copy link
Contributor

@nirda7 nirda7 commented May 27, 2025

No description provided.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pytest will automatically run every file that is named test_*.py , since this is a manual script example of how to run quantization, maybe add it under examples folder and rename the file name.
As well this repo is used by customers, if we add the script here we will need to maintain it constantly to be aligned with latest vllm interface. (@linoybu , @ulivne, what is your opinion?)

os.environ['EXPERIMENTAL_WEIGHT_SHARING'] = "0"
os.environ['PT_HPU_ENABLE_LAZY_COLLECTIVES'] = "true"
os.environ['PT_HPU_LAZY_MODE'] = "1"
#PT_HPUGRAPH_DISABLE_TENSOR_CACHE=true
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we want to remain with all the lines the are commented?
Maybe add some description of the available options in each argument and leave only one default LLM creation and the script user will edit it for himself.

end = time.time()
if is_fp8:
#if is_measure:
# llm.llm_engine.model_executor.shutdown_inc()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

delete?

@@ -0,0 +1,15 @@
{
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we should ship internal tests

@michalkuligowski michalkuligowski marked this pull request as draft June 27, 2025 12:26
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.

5 participants