Thanks to visit codestin.com
Credit goes to github.com

Skip to content

llama-cpp-python 0.3.8 with CUDA #2010

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.

Sign up for GitHub

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

Jump to bottom
Open
SeBL4RD opened this issue May 1, 2025 · 2 comments
Open

llama-cpp-python 0.3.8 with CUDA #2010

SeBL4RD opened this issue May 1, 2025 · 2 comments

Comments

@SeBL4RD
Copy link

SeBL4RD commented May 1, 2025

When will we have a recent version of llama-cpp-python, functional with CUDA, via pip?
It's a real nightmare to make it work any other way.
0.3.4 works with CUDA, but it doesn't take into account models like Qwen 3 quantified.

with kind regards
@JamePeng
Copy link

JamePeng commented May 2, 2025 •
edited
Loading

You can try to compile the new code I maintain here: https://github.com/JamePeng/llama-cpp-python, but I only pre-compiled the Windows and linux version based on the recent code

@m-from-space
Copy link

When will we have a recent version of llama-cpp-python, functional with CUDA, via pip?

I probably misunderstand your question, but I am using 0.3.9 with CUDA via pip. I can use Qwen3 related models as well (using Iquants for example or flash attention, is that what you are referring to?). Sorry if my comment is useless.

I built using...
CMAKE_ARGS="-DGGML_CUDA=on -DLLAVA_BUILD=off -DCMAKE_CUDA_ARCHITECTURES=native" FORCE_CMAKE=1 pip install llama-cpp-python --no-cache-dir --force-reinstall --upgrade

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@JamePeng @SeBL4RD @m-from-space