Freeze and/or exclude notebooks from sphinx build

[tms-bananaquit]

• Confirm that current pipeline runs the script versions to check for no error
• Freeze and/or do not execute very expensive examples (LFR, wilds)
  • Should we use small, representative datasets that are part of the repo for those?
• Motivating issue is the small example for pulling wilds one which has to load a big csv. If nothing else, makes build take a long time - if we keep going, we might end up running against restrictions on filesize/build-time for RTD. Pulling data from outside the repo also seems a little fragile.
  • **wilds_datasets.ipynb is currently excluded via exclude_patterns in conf.py to avoid this problem. That should be updated once this is fixed.