Run and Evaluate functions

Class the will run your source separation algorithm on a number of files then evaluate using the evaluation method of your choosing, all in one fell swoop.

nussl.evaluation.run_and_eval.run_and_evaluate(evaluation_object, evaluation_kwargs, separation_object, separation_kwargs, mixture_list, true_sources_list_of_lists)
Parameters:
  • evaluation_object (EvaluationBase) –
  • evaluation_kwargs (dict) –
  • separation_object (SeparationBase) –
  • separation_kwargs (dict) –
  • mixture_list
  • true_sources_list_of_lists

Returns:

nussl.evaluation.run_and_eval.run_and_eval_prf(separation_list, separation_kwargs, mixture_list, true_sources_list_of_lists, skip_errors=False, name_list=None, verbose=False)

This is a helper method to run a MaskSeparationBase-derived source separation algorithm on a list of mixtures (provided in mixture_list) Run and evaluate PrecisionRecallFScore for each :param separation_list: :param separation_kwargs: :param mixture_list: (list) List of AudioSignal objects that contain mixtures. :param true_sources_list_of_lists: (list) List of lists of AudioSignal objects that contain

true sources
Parameters:
  • skip_errors – (bool)
  • name_list – (list)

Returns: