After fixing for anomalies in the data (replacing null layers and layers with no data with a same-size matrix of zeros, here are the dimensions of the array for this run):
Each 3D tensor consists of 17 pads represented as length, with 24 timebins per pad, represented as breadth, and 6 layers, represented as depth. In run number 265377, there were 13858 tracklets that had transverse momenta between 2 and 3 GeV…