Installing LLAMA CUDA libraries and Python bindings ERROR

[kaisadhar]

While executing the command CMAKE_ARGS='-DLLAMA_CUBLAS=on' poetry run pip install --force-reinstall --no-cache-dir llama-cpp-python, I always get this error. Haven't found a solution for this problem. Ps: I am running this on WSL 2

[AlexPerkin]

Maybe this will help https://github.com/zylon-ai/private-gpt/issues/1584

[neofob]

@kaisadhar Recent versions of llama.cpp uses this flag -DLLAMA_CUDA=on instead of -DLLAMA_CUBLAS=on. So try to pass that to your poetry install... command.

Also, it requires package python3-dev. So with WSL, I guess you can do this:

wsl
sudo apt-get install -y python3-dev