As a DS user, I need to be able to configure notebook server environments with desired GPU resources, so I can improve performance for compute-intensive operations like model training.

Goal for this story:

• Investigate how GPUs are added to a cluster so we can determine how to get values for the # of supported GPUs. We want to provide users an option in the UI to select the desired # of GPUs for a NB server based on the # of GPUs available in the cluster. Ideally, we would auto populate the # of GPUs available for selection based on what's available. If that's not feasible, we may need to provide an option for RHODS admins to configure the # of GPUs available for selection.