divide multiple model reports into "quick" and "detailed", based on time needed

[nsfabina]

quick and dirty == order of seconds for readily available data

detailed == order of minutes for loading multiple models and doing predictions on many samples