Skip to content

Welcome to RuNNer!

The RuNNer Neural Network Energy Representation is a Fortran-based framework for the construction of Behler-Parrinello-type high-dimensional neural network potentials.

Version breaking changes in RuNNer 1.3

Be careful, with the new version we have introduced some breaking changes: Some keywords are now deprecated and need to be removed from existing input.nn files:

  • use_atom_charges
  • fixed_atom_charges


  • Construction of very flexible neural network potentials for the representation of potential energies in high-dimensional systems:
    • unlimited number of degrees of freedom (atoms).
    • training data can be obtained from arbitrary electronic structure methods and codes.
    • training using energies and forces.
    • periodic and non-periodic systems.
    • several types of descriptors for the atomic environments, including atom-centered symmetry functions, are available.
    • several types of activation functions are available
    • arbitrary topology of the atomic neural networks.
    • provides energies and analytic derivatives (forces and stress tensor).
  • Evaluation of the NNPs in junction with a MD or MC run.