This is material for a tutorial on machine learning presented at the SciDAC ImPACTS project meeting 2025.
- Log into a NERSC machine such as perlmutter in a terminal
- Clone this repository in your home directory.
An allocation has been created at NERSC for this tutorial. The tutorials are best run on jupyterhub at NERSC.
- Go to jupyter.nersc.gov
- Login using your NERSC credentials
- Use the trn009 allocation.
- Go to iris.nersc.gov and check trn009 as the default account. Then go to jupyter.nersc.gov and launch exclusive cpu or gpu
- Go straight to jupyter.nersc.gov and choose a configurable job and select trn009 as the account.
- In the configurable job, choose the time as 120 minutes and 1 GPU per node.
- Navigate to the mlaitutorial and choose a notebook to run.