Cannot install whl on centos7

[sdwcn]

Describe the bug

Cannot install whl on centos7, any workaround or newer whl might help?

Is there an existing issue for this?

• [X] I have searched the existing issues

Reproduction

pip install https://github.com/oobabooga/llama-cpp-python-cuBLAS-wheels/releases/download/textgen-webui/llama_cpp_python_cuda_tensorcores-0.2.26+cu121-cp311-cp311-manylinux_2_31_x86_64.whl
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
**ERROR: llama_cpp_python_cuda_tensorcores-0.2.26+cu121-cp311-cp311-manylinux_2_31_x86_64.whl is not a supported wheel on this platform.**

Screenshot

Logs

**ERROR: llama_cpp_python_cuda_tensorcores-0.2.26+cu121-cp311-cp311-manylinux_2_31_x86_64.whl is not a supported wheel on this platform.**

System Info

OS=centos 7, Linux 3.10.0-1160.105.1.el7.x86_64 x86_64 GNU/Linux, python = 3.11, CUDA 12.2 installed

[drc-pnnl]

I took out the wheel bits in the requirements.txt file. Specifically for lamma-cpp and replaced all the calls to potential wheels with

llama-cpp-python (CUDA, tensor cores)

llama-cpp-python

Now that requires of course that your system has a CUDA development environment and a recent compiler. to build llama-cpp-python during this install. My recollection is this is how things were 6 months or more ago that last time I updated the code.

[sdwcn]

@drc-pnnl I m using the matching url (in my case linux+x86+py3.11) from cuda tensor cores of requirements file to manually install the wheel, it gave me above error. just now i tried to compile on my own by "pip install llama-cpp-python==0.2.26" in conda env, it doesnt compile through with error:

const char* alpha = alphabet == alphabet::url_filename_safe  
**ERROR** ‘alphabet’ is not a class, namespace, or enumeration

[github-actions[bot]]

This issue has been closed due to inactivity for 2 months. If you believe it is still relevant, please leave a comment below. You can tag a developer in your comment.