I got access to DevCloud to do the coursework of DC to EDGE course. I connected via Jupyter and from home, I opened a new terminal in my browser.
import cv2 didn't work until I swapped steps 1 and 3. I also had to create the conda environment in a compute node to avoid exceeding resource limit on the login node.
Following the installation instructions in the course, I did the following:
(base) uxxxxxxxx@s001-n117:~$conda update conda
>> "EnvironmentNotWritableError: The current user does not have write permissions to the target environment."
(base) uxxxxxxxx@s001-n117:~$conda env create -f DC\ to\ Edge\ Course/Notebooks/environment.yml
(base) uxxxxxxxx@s001-n117:~$ conda activate tf_training
(tf_training) uxxxxxxx@s001-n117:~$ conda update conda
PackageNotInstalledError: Package is not installed in prefix.
package name: conda
yes, it is already running after doing the conda update as a last step. However, if this update is not done, I am not able to run the Notebook. So, I think it is not optional