代码:
import cntkn = cntk.minus([1, 2, 3], [4, 5, 6]).eval()print(n)问题:
如果在Python命令行中逐行运行代码，一切正常。
如果运行”Python main.py”，也一切正常。
但是如果在PyCharm或Visual Studio Code中运行代码，会出现错误，PyCharm的错误信息如下:
Traceback (most recent call last):File "D:\Apps\Anaconda3\envs\Python3.53\lib\site-packages\cntk\cntk_py.py", line 18, in swig_import_helper return importlib.import_module(mname)File "D:\Apps\Anaconda3\envs\Python3.53\lib\importlib\__init__.py", line 126, in import_module return _bootstrap._gcd_import(name[level:], package, level)File "<frozen importlib._bootstrap>", line 986, in _gcd_importFile "<frozen importlib._bootstrap>", line 969, in _find_and_loadFile "<frozen importlib._bootstrap>", line 956, in _find_and_load_unlockedImportError: No module named 'cntk._cntk_py'During handling of the above exception, another exception occurred:Traceback (most recent call last): File "D:/Dev/machine_learning/Test_CNTK/main.py", line 1, in <module> import cntk File "D:\Apps\Anaconda3\envs\Python3.53\lib\site-packages\cntk\__init__.py", line 11, in <module> from .core import * File "D:\Apps\Anaconda3\envs\Python3.53\lib\site-packages\cntk\core.py", line 10, in <module> from . import cntk_py File "D:\Apps\Anaconda3\envs\Python3.53\lib\site-packages\cntk\cntk_py.py", line 21, in <module> _cntk_py = swig_import_helper() File "D:\Apps\Anaconda3\envs\Python3.53\lib\site-packages\cntk\cntk_py.py", line 20, in swig_import_helper return importlib.import_module('_cntk_py') File "D:\Apps\Anaconda3\envs\Python3.53\lib\importlib\__init__.py", line 126, in import_module return _bootstrap._gcd_import(name[level:], package, level)ImportError: DLL load failed: The specified module could not be found.环境:
- Anaconda: 4.3.1
- Python: 3.5.3 (在anaconda中的虚拟环境)
- CNTK: 2.0beta12 适用于Windows amd64 GPU
- PyCharm: 2016.3.2
- Visual Studio Code: 1.10.2
回答：
已解决
因为它找不到CNTK相关的DLL文件（例如CNTKLibrary-2.0.dll, EvalDll.dll），这些DLL文件在安装CNTK后与”python.exe”位于同一目录中。
有两种解决方案:
（假设PYTHONDIR是python.exe所在的目录）
-
更改系统环境变量PATH以包含PYTHONDIR
-
在”import cntk”之前添加以下代码:
import os; os.environ[‘PATH’] = PYTHONDIR + ‘;’ + os.environ[‘PATH’]
