Failed to load shared library error: undefined symbol _ZNSt15__exception_ptr13exception_ptr9_M_addrefEv

[kausthub-kannan]

Found a similar issue but I haven't set any CXX env variable while installing llama-cpp-python. I feel it's gcc version issue (currently the deployment system is on gcc v10 but works on my PC which has gcc v11)

OS: Debian
llama-cpp-python Version: 0.2.87
python version: 3.9.2
gcc/g++ version: 10.2.1

Here is the log when I try to import lllama_cpp:

Python 3.9.2 (default, Feb 28 2021, 17:03:44)
[GCC 10.2.1 20210110] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import llama_cpp
Traceback (most recent call last):
File "/usr/local/lib/python3.9/dist-packages/llama_cpp/llama_cpp.py", line 75, in _load_shared_library
return ctypes.CDLL(str(_lib_path), **cdll_args)  # type: ignore
File "/usr/lib/python3.9/ctypes/__init__.py", line 374, in __init__
self._handle = _dlopen(self._name, mode)
OSError: /usr/local/lib/python3.9/dist-packages/llama_cpp/lib/libllama.so: undefined symbol: _ZNSt15__exception_ptr13exception_ptr9_M_addrefEv

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python3.9/dist-packages/llama_cpp/__init__.py", line 1, in <module>
from .llama_cpp import *
File "/usr/local/lib/python3.9/dist-packages/llama_cpp/llama_cpp.py", line 88, in <module>
_lib = _load_shared_library(_lib_base_name)
File "/usr/local/lib/python3.9/dist-packages/llama_cpp/llama_cpp.py", line 77, in _load_shared_library
raise RuntimeError(f"Failed to load shared library '{_lib_path}': {e}")
RuntimeError: Failed to load shared library '/usr/local/lib/python3.9/dist-packages/llama_cpp/lib/libllama.so': /usr/local/lib/python3.9/dist-packages/llama_cpp/lib/libllama.so: undefined symbol: _ZNSt15__exception_ptr13exception_ptr9_M_addrefEv

I am trying to run on CPU only device (plus a low end device) and this the installation command which I used:

pip install llama-cpp-python==0.2.87 --force-reinstall --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu

Also it would be helpful to know if I am installing the package for CPU only device (as I need to avoid excess use of disk space)

[kausthub-kannan]

Was able to fix this error for version 0.2.78. I was trying run llama cpp on a VM. A similar issue was faced by few when they were trying to build llama cpp (without binding) from source. While installing llama-cpp-python, I had to disable llama native.

 CMAKE_ARGS="-DLLAMA_NATIVE=off" pip install llama-cpp-python==0.2.78 --force-reinstall --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu