Skip to content

@JamesPeng, Thank you for this work around to enable qwen 3.6. I would like to say... #141

@patrikpatrik

Description

@patrikpatrik

Instead of windows powershell option under Installation Configuration -> Environment Variables -> can you also add right below it:

#Windows Command Prompt
set CMAKE_ARGS = "-DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS"

that's for those who just run cmd + R for command prompt.
and if in .\.venv\scripts\python.exe environment for desktop you would run:
C:\Users\<username>\ComfyUI-Installs\ComfyUI\.venv\Scripts\python.exe -m pip install "\path\to\llama_cpp_python-0.3.40+cu131-cp313-cp313-win_amd64.whl"

Because when using this line:
python.exe -m pip install "llama-cpp-python @ git+https://github.com/JamePeng/llama-cpp-python.git"
There's some Error where it cannot download files. I'm not sure why it keeps doing that consistently.

Then do a sanity check after wheel installation by:
python.exe -c "from llama_cpp import Llama; print('llama-cpp import OK')"

But thank you for this. I finally got CUDA installed and the correct wheel for Qwen 3.6 on ComfyUI! 👏👍

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions