Data conversion from root files to h5 files is needed to run the ML model training: run python datasaver.py

• Uproot
• Pandas

To run use python training_scheme.py X where X is:

• NN for Neural Network training

• NNOptimiser for hyperparameter optimisation

• GBDT for GBDT training

• GBDTOptimiser for hyperparameter optimisation

For NN:

• Tensorflow 2.0
• Tensorflow_model_optimization (for pruning)
• Keras
• qkeras (for quantised aware training) For GBDT:
• xgboost
• joblib (saving models in pickle format)

For Both:

• Pandas
• Sklearn (for metrics and utility functions)
• Comet_ml (for parameter logging, will also need an account and API key to log metrics)
• yaml (for parameter file parsing)