Getting a numpy import error when numpy is installed

[AndreTInfante]

No description provided.

[AndreTInfante]

Sorry, accidentally submitted before filling this out. I'm getting this error on windows 10, python 3.5.

when I try to run the ball balancing example project. I initially thought it was because I had the wrong version of numpy installed, but I reverted to 1.11.0, and it's still throwing the issue. Any thoughts on what might be up?

[AndreTInfante]

For reference:

(tensorflowcpu) C:\Users\aticp\Downloads\ml-agents-master\ml-agents-master\python>pip3 install numpy
Requirement already satisfied (use --upgrade to upgrade): numpy in c:\users\aticp\appdata\local\programs\python\python35\lib\site-packages

[vincentpierre]

I think there is something wrong with jupyter, not python.

• Is the Kernel of the Jupyter notebook set to Python3 or Python2 ?
• Do you have multiple versions of Python installed ? If yes, jupyter must use the same python as pip.
• What happens if you open Python3 from the command line and try to import numpy?
• I think you are using a virtual environment called tensorflowcpu, are you sure you launched jupyter notebookfrom the same virtual environment ?

Let us know if you find a solution, we are interested in knowing about this issue.

[AndreTInfante]

The tensorflowcpu environment is made using anaconda, to try to simplify things. I definitely launched jupyter notebook from within the same environment. I'll try setting things up outside of anaconda as well, and see if I have any more luck.

mporting numpy in python seems to work (or at least throws no errors.

The kernel.json file shows

{
"language": "python",
"display_name": "Python 3",
"argv": [
"python",
"-m",
"ipykernel_launcher",
"-f",
"{connection_file}"
]
}

Which looks correct to me, since I'm trying to use python 3.5. Anything else I should check?

Thanks!

[AndreTInfante]

Ah, interesting! Doing it outside of anaconda seems to have fixed the issue. Weird. Thank you for your help!

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.