为什么当我运行代码时，我可以看到pip-list sklearn，但不能在Jupyter中看到

Question

我目前正在使用Jupyter编写我的第一个程序。当我运行一个单元格时，我会收到以下

<ipython-input-6-7a7db0de4539> in <module>
1 try:
----> 2     from sklearn.datasets import fetch_openml
3     mnist = fetch_openml('mnist_784', version=1, cache=True)
c:usersjamesappdatalocalprogramspythonpython37libsite-packagessklearn__init__.py in <module>
63     from . import __check_build
---> 64     from .base import clone
65     from .utils._show_versions import show_versions
c:usersjamesappdatalocalprogramspythonpython37libsite-packagessklearnbase.py in <module>
12 from .externals import six
---> 13 from .utils.fixes import signature
14 from . import __version__
c:usersjamesappdatalocalprogramspythonpython37libsite-packagessklearnutils__init__.py in <module>
15 from ..exceptions import DataConversionWarning
---> 16 from .fixes import _Sequence as Sequence
17 from .deprecation import deprecated
c:usersjamesappdatalocalprogramspythonpython37libsite-packagessklearnutilsfixes.py in <module>
84 else:
---> 85     from scipy.special import boxcox  # noqa
86 
c:usersjamesappdatalocalprogramspythonpython37libsite-packagesscipyspecial__init__.py in <module>
640 
--> 641 from ._ufuncs import *
642 
ImportError: DLL load failed: The specified module could not be found.

我使用pip安装来获得sklearn，当在cmd中执行pip列表时，它就在那里。然而，当我尝试导入sklearn时，我会遇到一个错误。

我已经检查了import sys和sys.path，这给了我：

'C:\Users\James\Documents\Machine Learning Comp Vision\Project 1',
'C:\Users\James\Documents\Machine Learning Comp Vision\Project 1',
'c:\users\james\appdata\local\programs\python\python37\python37.zip',
'c:\users\james\appdata\local\programs\python\python37\DLLs',
'c:\users\james\appdata\local\programs\python\python37\lib',
'c:\users\james\appdata\local\programs\python\python37',
'',
'c:\users\james\appdata\local\programs\python\python37\lib\site-packages',
'c:\users\james\appdata\local\programs\python\python37\lib\site-packages\IPython\extensions',
'C:\Users\James\.ipython']

我的所有模块都位于C:UsersJamesAppDataLocalProgramsPythonPython37libsite-packages中我的文件url中的大写字母是造成这种情况的原因吗？还是与路径本身有关？

Answer 1

如果您已经安装了numpy和scipy，请执行以下操作：

pip install -U scikit-learn

否则：

conda install scikit-learn

最后检查更新：

conda update pip

Answer 2

使用Jupyter笔记本内部的pip

!{sys.executable} -m pip install -U scikit-learn

这将使用sys.executable为当前内核运行正确的pip，而不是为其他Python安装运行。