09-21-2018 01:58 AM
OS: Windows 10
Python 3.6.1 (v3.6.1:69c0db5, Mar 21 2017, 18:41:36) [MSC v.1900 64 bit (AMD64)] on win32
After performing clean install of native Python (without numpy, pandas, and etc.), inside specific virtual environment, and installing of intel-sklearn with:
pip install intel-scikit-learn subsequent error has appeared:
Subsequent error has appeared, making usage of the packages impossible:
Importing the multiarray numpy extension module failed. Most
likely you are trying to import a failed build of numpy.
If you're working with a numpy git repo, try `git clean -xdf` (removes all
files not under version control). Otherwise reinstall numpy.
Original error was: DLL load failed: The specified module could not be found.
09-26-2018 09:27 PM
I investigated this further and was able to root-cause the underlying problem - the issue lies with how virtualenv patches PATH variable once the activation script is run. The python virtualenv module generates a template-based activation script which doesn't account for extra set of directories that ought to be added to PATH for installed binaries to work correctly. Proposed work-around for current behavior of virtualenv is to ensure that activation script adds those required directories to PATH.
1. Create a venv called "test"
2. Open ".\test\Scripts\activate.bat" in an editor
Change this line :
3. Activate "test" venv
Inside this venv, you would be able to install and use our packages. And since the change is at the environment-activation-script level, all the required set of directories would be set to PATH, and would hence be a one-time setup for that venv.
Hope this helps!
09-27-2018 02:21 AM
Dear Rohit, thank you very much for your support. Now, after adjusting the path, everything is working correctly.
I hope there will be a benefit to others if similar problem is encountered.
Thank you once again.