Windows ModuleNotFoundError: No module named 'torch'

vision
1

I am using a Windows environment with Conda (no cuda installed). I am able to run test.py in my jupyter notebook(without any toch error) but it as child process forking issues.

But I am unable to run my test.py on my conda prompt as it throws the below error when I run:

ModuleNotFoundError: No module named 'torch'

Python Version

(base) C:\Users\MAHE\Modified Unet3D Master -TestRun>python --version
Python 3.6.4 :: Anaconda, Inc.

Corresponding torch installed

Downloading and Extracting Packages
torchvision-cpu-0.2. | 37 KB     | ############################################################################################################################################ | 100%
pytorch-cpu-1.0.1    | 58.8 MB   | ############################################################################################################################################ | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done

I have updated conda and also run

conda install mkl=2018

as it worked for some people. Nothing seems to be working.

Please help.

2

Okay, I resolved this by using this script to check

and deleting my other python version (even though I had no variables pointing there).

Also py test.py does not work anymore.

I need to do python test.py.

Strange.