Skip to content

Add optional ML based e/p workflow

This adds an optional workflow, not to be used in production, that predicts the e/p for a track using a ML model. The main changes are the addition of a new tool, PFEnergyPredictorTool written by @abarton , along with associated toggles to enable usage of this via a python flag. I tested on 2000 ttbar events that the output looks the same with physics validation. RunWorkFlowTests_run3.py indicated no frozen violating tier0 changes. The above is as expected, given this workflow is not enabled by default.

Merge request reports