Using a Jupyter Notebook on computing nodes

Step 1

Log into the cluster, e.g. with NX NoMachine.

Step 2

Create a SLURM-Script that initializes a Jupyter Notebook instance. If you prefer to do so, you can directly activate your conda environment in the SLURM-script. An example for starting the Jupyter Notebook:

#!/bin/bash
#SBATCH --gres=gpu:V100:1                 # Request 1 GPU
#SBATCH --mem=4096           
#SBATCH -t 30
#SBATCH -e job-%j.err	# Inside this file you will find the url

jupyter-notebook --no-browser --port=8080 --ip "*"

Step 3

You should find the url address inside the error file. The file will be created in the same directory as your .sh file - if you can't seem to find the file via your file explorer, then "cd" in the shell towards the path and try the "ls" command to list hidden files. You can open the file inside the command line interface with an editor like nano or Vim.