Support inference on multiple models simultaneously

[drewoldag]

An open question for the science users - would it beneficial to be able to define a set of models to use for inference?

Given a collection of models that must all be used to predict results for a given data set we could either run the data through each model individually (or in manually in parallel with a collection of scripts) or we could parallelize this within FIBAD such that the input data is accessed once, and processed in parallel by many models.