class atomistic.Polarizability(*args: Any, **kwargs: Any)[source]

Predicts polarizability tensor using tensor rank factorization. This requires an equivariant representation, e.g. PaiNN, that provides both scalar and vectorial features.


  • n_in – input dimension of representation

  • n_hidden – size of hidden layers. If an integer, same number of node is used for all hidden layers resulting in a rectangular network. If None, the number of neurons is divided by two after each layer starting n_in resulting in a pyramidal network.

  • n_layers – number of layers.

  • activation – activation function

  • polarizability_key – the key under which the predicted polarizability will be stored