Skip to content

Adding the new DIPZ architecture to the HLT code

Mohammed Aboelela requested to merge maboelel/athena:adding_DIPZ_v2 into 23.0

Modified the required three files in the listed packages below to add a link to the DIPZ lwtnn that exists in the dev area, and save the outputs of the neural network to jets (assigning mean_z and sigma_z to each jet) inside the events accepted by HLT. DIPZ is an implemented neural network (that uses DIPS taggers) whose purpose is to regress the jet's vertex z-position.

Successfully ran and cross-checked these tests:

1- test_trig_data_v1Dev_build.py inside runTrigART.py

2- test_trigAna_RDOtoAOD_v1Dev_build.py

The list of packages touched by the changes:

Trigger/TrigSteer/TrigOutputHandling

Trigger/TriggerCommon/TrigEDMConfig

Trigger/TriggerCommon/TriggerMenuMT

Merge request reports