The compiled libllama.dll could NOT be found.

[ikionchiu]

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](https://github.com/abetlen/llama-cpp-python/blob/main/README.md).
• [x] I [searched using keywords relevant to my issue](https://docs.github.com/en/issues/tracking-your-work-with-issues/filtering-and-searching-issues-and-pull-requests) to make sure that I am creating a new issue that is not already open (or closed).
• [x] I reviewed the [Discussions](https://github.com/abetlen/llama-cpp-python/discussions), and have a new bug or useful enhancement to share.

Expected Behavior

• I hope the compiled libllama.dll file can be used normally in other directories as well.
• I hope the following Python script can run correctly.

Current Behavior

• The currently compiled libllama.dll file reports errors and cannot be used normally in other directories.
• The following Python script can NOT run correctly.

Environment and Context

Physical hardware i am using:

• Operating System: Windows 10 Pro 21H2 19044.2965 64bit x64
• CPU: 11th Gen Intel(R) Core(TM) i5-1135G7 @ 2.40GHz 2.42 GHz

SDK version(The CMake and MinGW folders were copied from CLion and added separately to the environment variables):

• python: 3.10.9
• llama_cpp_python: 0.2.69
• pywin32-ctypes: 0.2.2
• cmake: 3.28.1 (CMake; JetBrains IDE bundle; build 16)
• g++: 13.1.0
• gcc: 13.1.0

Steps to Reproduce

1. step 1: Create env folder and placed in the same directory level as llama-cpp-python.

2. step 2: Copy the CMake and MinGW folders from CLion directory and add to the env folder.

3. step 3: Add CMake and MinGW to the system environment variables.(cmake: ***\env\cmake\win\x64\bin; mingw:***\env\mingw\bin)

4. step 4: cd llama-cpp-python

5. step 5: mkdir build

6. step 6: cd build

7. step 7: cmake .. -DBUILD_SHARED_LIBS=ON -G "MinGW Makefiles"

8. step 8: cmake --build . --config Release

9. step 9: Then, I can find the libllama.dll file in the directory ***\llama-cpp-python\build\bin\.

10. step 10: Then, I copied the generated libllama.dll to another directory for testing, but the test failed. I pinpointed the issue to the following code:

    import ctypes
    import os
    cdll_args = dict()
    cdll_args["winmode"] = ctypes.RTLD_GLOBAL
    os.add_dll_directory('D:/temp/test/llama_cpp')
    _lib_path = 'D:/temp/test/llama_cpp/libllama.dll'
    ctypes.CDLL(str(_lib_path), **cdll_args)

Failure Logs

The ERROR log as following:

Traceback (most recent call last):
  File "D:\temp\test.py", line 25, in <module>
    ctypes.CDLL(str(_lib_path), **cdll_args)
  File "C:\Users\testuser\AppData\Local\Programs\Python\Python310\lib\ctypes\__init__.py", line 374, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: [WinError 127] The specified procedure could not be found

[SavorSauc3]

Have you tried checking the directory that the llama_cpp is in? On linux machines the name of the .so file is libllama.so, but on windows machines where the file is a .dll, the names are llama.dll and llava.dll. You could try the following code and check if it works: _lib_path = 'D:/temp/test/llama_cpp/llama.dll'