In order to ease up the usage of Two!Ears model, the training of the
probabilistic models used for the knowledge sources may be skipped by the user.
This is a collection of data created during a-priori training of the models.
The data for the respective knowledge source is stored inside
learned_models/<name of knowledge source class>. Documentation of the
training data is available inside the description of the individual
:ref:`sec-knowledge-sources`.