Published February 17, 2020 | Version 2.0
Software Open


  • 1. CERN
  • 2. Nat. Cent. for Sci. Res. Demokritos
  • 3. Vrije Universiteit Brussel (VUB)
  • 4. Helsinki Institute of Physics (HIP)
  • 5. Univ. of California Santa Barbara


DeepJetCore: Package for training and evaluation of deep neural networks for HEP

This package provides the basic functions for out-of-memory training, resampling, and basic evaluation. For simple use cases, only two files need to be adapted: the actual training data structures, describing how to fill numpy arrays from root trees, and the DNN model itself. Both must be defined in an additional user package. Please refer to the Section 'Usage' for more information.


Files (5.2 MB)

Name Size Download all
5.2 MB Preview Download