Tensorflow and torch not found

post by trm on Feb 13, 2023

I installed lambdastack on fresh Ubuntu 22.04 following the instructions here: Lambda Stack: an AI software stack that's always up-to-date

I.e. I ran

wget -nv -O- https://lambdalabs.com/install-lambda-stack.sh | sh -
sudo reboot

After reboot python3 wasn’t able to import tensorflow or torch. So I reinstalled following instructions here:

https://support.lambdalabs.com/hc/en-us/articles/9060760997005-Remove-and-reinstall-Lambda-Stack

After reboot still no dice. This is my output:

~$ python3
Python 3.10.9 (main, Jan 11 2023, 15:21:40) [GCC 11.2.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.

>>> import tensorflow as tf
>>> Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ModuleNotFoundError: No module named ‘tensorflow’

>>> import torch
>>> Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ModuleNotFoundError: No module named ‘torch’

I can see tensorflow and torch in /usr/lib/python3/dist-packages.

How can I resolve this issue?

post by markd on Jun 13, 2023

A common issue is if other things are installed in ~/.local or /usr/local.

Doing the following would likely provide the information of what is seen:

$ pip -v list | egrep “torch|tensorflow”

So if you are inside a virtual environment (Anaconda or python venv) it would ‘hide’ the system installed packages.

To version your CUDA you can use a virtual environment or docker container.

I have a number of examples at:

We should have better write up soon at ‘http://docs.lambdalabs.com/’