get_best_model_and_data

behavenet.fitting.utils.get_best_model_and_data(hparams, Model=None, load_data=True, version='best', data_kwargs=None)[source]

Load the best model (and data) defined by hparams out of all available test-tube versions.

Parameters:

  • hparams (dict) – needs to contain enough information to specify both a model and the associated data

  • Model (behavenet.models object, optional) – model type

  • load_data (bool, optional) – if False then data generator is not returned

  • version (str or int, optional) – can be ‘best’ to load best model

  • data_kwargs (dict, optional) – additional kwargs for data generator

Returns:

  • model (behavenet.models object)

  • data generator (ConcatSessionsGenerator object or NoneType)

Return type:

tuple