Run first the following:

source Setup.sh

Use the following image: gitlab-registry.cern.ch/ai-ml/kubeflow_images/pytorch-notebook-gpu-1.8.1:v0.6.1-30

Use the CreateH5files.py script to produce H5 files with the information needed for the training.

How to run?

Edit the datasets dictionary if needed before running.

python CreateH5files.py

You can use the CHeckH5file.py script to make sure a given H5 file make sense (content is right).

You can also use it as a reference on how to open a H5 file and get all the data.

You can use MakePlots.py to make some plots from a given H5 file (currently only making a single plot comparing HT between events w/ and w/o quark-jets).