ImportError: DLL load failed while importing segmented_maxsim_cpp:

`C:\Users\jam\AppData\Local\anaconda3\envs\python39\python.exe C:\Users\jam\AppData\Roaming\JetBrains\IntelliJIdea2021.2\plugins\python\helpers\pydev\pydevd.py --multiproc --qt-support=auto --client 127.0.0.1 --port 51622 --file C:\Users\jam\AppData\Roaming\JetBrains\IntelliJIdea2021.2\plugins\python\helpers\pycharm\_jb_pytest_runner.py --path D:/SourceCodes/neuron_llm/llama_index/testing/test_colbert_v2.py
Testing started at 14:10 ...
Connected to pydev debugger (build 212.5712.43)
C:\Users\jam\AppData\Roaming\JetBrains\IntelliJIdea2021.2\plugins\python\helpers\pycharm\_jb_pytest_runner.py:33: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
Launching pytest with arguments D:/SourceCodes/neuron_llm/llama_index/testing/test_colbert_v2.py --no-header --no-summary -q in D:\SourceCodes\neuron_llm\llama_index\testing
elif version.LooseVersion(pytest.__version__) >= version.LooseVersion("6.0"):

============================= test session starts =============================
collecting ... [Mar 14, 14:10:24] Loading segmented_maxsim_cpp extension (set COLBERT_LOAD_TORCH_EXTENSION_VERBOSE=True for more info)...
1.11.1.git.kitware.jobserver-1
Using C:\Users\jam\AppData\Local\torch_extensions\torch_extensions\Cache\py39_cpu as PyTorch extensions root...
Emitting ninja build file C:\Users\jam\AppData\Local\torch_extensions\torch_extensions\Cache\py39_cpu\segmented_maxsim_cpp\build.ninja...
Building extension module segmented_maxsim_cpp...
Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
Loading extension module segmented_maxsim_cpp...

test_colbert_v2.py:None (test_colbert_v2.py)
ImportError while importing test module 'D:\SourceCodes\neuron_llm\llama_index\testing\test_colbert_v2.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\_pytest\python.py:617: in _importtestmodule
mod = import_path(self.path, mode=importmode, root=self.config.rootpath)
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\_pytest\pathlib.py:567: in import_path
importlib.import_module(module_name)
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\importlib\__init__.py:127: in import_module
return _bootstrap._gcd_import(name[level:], package, level)
:1030: in _gcd_import
???
:1007: in _find_and_load
???
:986: in _find_and_load_unlocked
???
:680: in _load_unlocked
???
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\_pytest\assertion\rewrite.py:186: in exec_module
exec(co, module.__dict__)
test_colbert_v2.py:41: in
RAG = RAGPretrainedModel.from_pretrained("D:\\SourceCodes\\neuron_llm\\llama_index\\testing\\colbertv2")
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\ragatouille\RAGPretrainedModel.py:71: in from_pretrained
instance.model = ColBERT(
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\ragatouille\models\colbert.py:80: in __init__
self.inference_ckpt = Checkpoint(
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\colbert\modeling\checkpoint.py:19: in __init__
super().__init__(name, colbert_config)
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\colbert\modeling\colbert.py:23: in __init__
ColBERT.try_load_torch_extensions(self.use_gpu)
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\colbert\modeling\colbert.py:38: in try_load_torch_extensions
segmented_maxsim_cpp = load(
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\torch\utils\cpp_extension.py:1308: in load
return _jit_compile(
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\torch\utils\cpp_extension.py:1736: in _jit_compile
return _import_module_from_library(name, build_directory, is_python_module)
C:\Users\jam\AppData\Local\anaconda3\envs\python39\lib\site-packages\torch\utils\cpp_extension.py:2136: in _import_module_from_library
module = importlib.util.module_from_spec(spec)
:565: in module_from_spec
???
:1173: in create_module
???
:228: in _call_with_frames_removed
???
E ImportError: DLL load failed while importing segmented_maxsim_cpp: 找不到指定的模块。
collected 0 items / 1 error

!!!!!!!!!!!!!!!!!!! Interrupted: 1 error during collection !!!!!!!!!!!!!!!!!!!!
=================== 2 warnings, 1 error in 61.02s (0:01:01) ===================

Process finished with exit code 2
`

Begin with testing/test_colbert_v2.py and its RAGPretrainedModel.from_pretrained call, then follow ColBERT.try_load_torch_extensions and torch.utils.cpp_extension.load in the traceback. Reproduce under the reported Windows/Python 3.9 setup; done means test collection imports segmented_maxsim_cpp without the DLL error.

Written by the indexing model from the issue text.