How to install an EDITABLE version of llama-cpp-python to multiple conda envs?

[devzzzero]

Hi, I installed llama-cpp-python to a miniconda env (with some errors) using
CUDACXX=/usr/local/cuda-12.3/bin/nvcc CMAKE_ARGS="-DLLAMA_CUDA=on -DCMAKE_CUDA_ARCHITECTURES=native" FORCE_CMAKE=1 pip install -e . --no-cache-dir --force-reinstall --upgrade --verbose

Now. this process takes about 20 minutes.. .

I'd like to "clone this" to other conda envs, but without going through the hassles of recompiling. These other envs have slightly different versions of other packages. I would like to install an editable version of llama-cpp-python to these.

Ideally, I'd like to dump the llama.cpp .so binaries and the like to a single location (say /pkgs/llama-cpp-python/) and have all my miniconda envs use the same EDITABLE llama-cpp-python directory for the python bits)

Because each conda env is ultimately using the same nvidia tools (in /usr/local/cuda/bin) I do not need to recompile the binary .so files for each venv. It's a waste of resources, and time)

How do I go about doing this? Is this even possible? The key thing here is that I'd like to share the same .so files for each miniconda env.

Thank you.