pY/pX & pZ/pX

Implementation in clusterization code

Got the code implemented with the selection for the i x j x k grid for the neural network and the neural network application for 1 class classification. Now trying to get the code to compile with the OnnxRuntime library. Current implementation is probably slow and memory inefficient, but it's a start.

Training data

Submitted 5 x 50 Ev. 50 kHz PbPb simulations: Finished yesterday night, will downsample along pT (Tsallis distribution), row and pad (histogram) now and retrain network. -> Checked the training data size: Was previously only 500k points, so expect a good improvement for regression quality with more data!

ONNX GPU implementation

Tried some more cmake settings, but no success yet...