We refactored the azure pipeline a bit in previous sprints to handle non-convergence better and to separate model runs from post-processing. We tested it on a few states and dates, but we should rerun it on all states and forecast dates, and then pull those outputs from blob storage onto local
Requirements
[ ] test out a small azure run, look at outputs, make sure they are what we want
Goal
We refactored the azure pipeline a bit in previous sprints to handle non-convergence better and to separate model runs from post-processing. We tested it on a few states and dates, but we should rerun it on all states and forecast dates, and then pull those outputs from blob storage onto local
Requirements