Skip to content
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

[Bug]: When using the latest 0.6.3, No module named 'vllm._version' appears #9421

Open
1 task done
lixiaolx opened this issue Oct 16, 2024 · 12 comments
Open
1 task done
Labels
bug Something isn't working

Comments

@lixiaolx
Copy link

Your current environment

The output of `python collect_env.py`
Your output of `python collect_env.py` here

Model Input Dumps

No response

🐛 Describe the bug

Start service: vllm serve /models/huggingface.co/meta-llama/Llama-3-8b-hf/
A warning appears:
/usr/local/lib/python3.10/dist-packages/vllm/connections.py:8: RuntimeWarning: Failed to read commit hash: No module named 'vllm._version' from vllm.version import __version__ as VLLM_VERSIO

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.
@lixiaolx lixiaolx added the bug Something isn't working label Oct 16, 2024
@leoguillaume
Copy link

leoguillaume commented Oct 16, 2024

Same issue with mistral-nemo-instruct-2407 and llama31-8b-instruct.

@simon-mo
Copy link
Collaborator

I will release a new version once this is fixed... @dtrifiro I think this is due to the fact I hard coded version in setup.py's so no _version.py is generated. Is that correct?

def get_vllm_version() -> str:
    return "0.6.3." # i did this, which skipped _version.py writing. 
    version = get_version(
        write_to="vllm/_version.py",  # TODO: move this to pyproject.toml
    )

@lixiaolx
Copy link
Author

I will release a new version once this is fixed... @dtrifiro I think this is due to the fact I hard coded version in setup.py's so no _version.py is generated. Is that correct?

def get_vllm_version() -> str:
    return "0.6.3." # i did this, which skipped _version.py writing. 
    version = get_version(
        write_to="vllm/_version.py",  # TODO: move this to pyproject.toml
    )

I don’t know if the following is the same problem?
When I tried to compile locally, I found that the final wheel package, or pip Install -e . showed that the version of vllm was not 0.6.3, but vllm-0.1.3.dev2745+g79799263.d20241016.cu123-cp310-cp310- linux_x86_64.whl

@CassieFeng
Copy link

same issue with glm-4v-9b
cuda 12.1 vllm=0.6.3

@ganisback
Copy link

+1

@wsiwei
Copy link

wsiwei commented Oct 17, 2024

same issue with qwen2-vl-instruct
cuda12.2 vllm=0.6.3

@matian388
Copy link

+1

@dtrifiro
Copy link
Contributor

@lixiaolx you can fetch tags using git fetch --tags and then the verison installed with pip install -e . will report the correct version. You can also check that it's correct by running pip install setuptools-scm; python -m setuptools_scm

@simon-mo yes, that's the reason why this is happening.

@simon-mo
Copy link
Collaborator

I will make a patch release once #9375 merged

@JieShenAI
Copy link

I use the following command to solve this issue:

pip install "vllm>=0.4.3,<0.6.4"

@bbotte
Copy link

bbotte commented Jan 9, 2025

how is now,
other version tip error:
error: subprocess-exited-with-error
× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> [17 lines of output]

@ArlanCooper
Copy link

+1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests