Skip to main content
Print

Conda Environments in jupyter notebooks

You can create a custom Conda environment on the cluster and then allow the Jupyter notebook to use that environment in open on demand. The steps to take to accomplish that are:

  1. Log into the cluster using a terminal/PuTTy (https://arc.uc.edu/getting-connected/)
  2. once Logged in, load the anaconda module using the command
    • module load anaconda/3
  3. Create a conda environment
    • conda create --name jupyter_env
  4. Install which ever modules you would like into that environment
    • source activate jupyter_env
    • conda install pandas tensorflow
  5. Install the ipykernel module
    • conda install ipykernel
  6. Link the environment to Jupyter
    • python -m ipykernel install --user --name jupyter_env --display-name "Python (test_env)"
  7. Once that is done you can start a new Jupyter instance on the ondemand webpage. You should see the new environment when you click on the “new” dropdown menu in Jupyter (see image below)
Table of Contents