llama_cpp installs into base instead of Conda for M1 installation

[nivibilla]

Prerequisites

Please answer the following questions for yourself before submitting an issue.

• [x] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
• [x] I carefully followed the README.md.
• [x] I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• [x] I reviewed the Discussions, and have a new bug or useful enhancement to share.

Expected Behavior

Install into the Conda library

Current Behavior

Installs into base python

Environment and Context

Please provide detailed information about your computer setup. This is important in case the issue is not reproducible except for under certain specific conditions.

• Physical (or virtual) hardware you are using, e.g. for Linux:

M1 Mac

• Operating System, e.g. for Linux:

MacOS

• SDK version, e.g. for Linux:

python 3.9

Steps to Reproduce

Please provide detailed steps for reproducing the issue. We are not sitting in front of your screen, so the more detail the better.

follow instruction for m1 Mac but do it inside Conda env

[abetlen]

Can you install into the conda env with verbose output and paste that here? Should give me an idea of what's going wrong.

[ianscrivener]

@nivibilla check out the MacOS install instructions here: https://github.com/abetlen/llama-cpp-python/blob/main/docs/macos_install.md

did you skip the conda activate llama step?

[nivibilla]

Hey, no Im pretty sure I built it inside the conda env. I haven't got round to testing this again. Will try do it tomorrow

[K-Mistele]

+1 on this, copied/pasted the install commands from instructions and have uninstalled & reinstalled it multiple times in the llama env, but passing --n_gpu_layers -1 (-1 = all layers), or any other number of layers is SLOWER than CPU inference until I switch back to the base env, at which point it's super fast. Happy to make a loom demonstrating this :)

Macbook pro M2 Max here, latest version of llama-cpp-python, which as of today is 0.2.29