We are now re-indexing data backwards from epoch start
We are now re-indexing data backwards from epoch start in order to avoid unnecessary gaps, plus controlling output size from the schedullerr based on the epoch start, horizon, and forward prediction values, allowing for 1hour of forward predictions