Use predictions(fromBatch:) to do inference on multiple tiles at once (and increase performance?)

[mshamash]

https://developer.apple.com/documentation/coreml/mlmodel/3088750-predictions

https://developer.apple.com/documentation/coreml/mlbatchprovider

Doing inference on multiple tiles at once can improve performance for plaque detection. Consider doing so with BatchRequests and sending multiple inputs to the model directly.