mlrl.testbed.experiments.output.evaluation.measurements module

Author: Michael Rapp (michael.rapp.ml@gmail.com)

Provides classes for keeping track of several measurements according to different measures.

class mlrl.testbed.experiments.output.evaluation.measurements.Measurements(num_values_per_measure: int)

Bases: object

Keeps track of values that correspond to different measures.

average_by_measure(measure: Measure) → Tuple[float, float]

Returns an average and a corresponding standard deviation for a given measure. The average is calculated as the arithmetic mean of all values that have been tracked for the measure.

Parameters:

measure – The measure

Returns:

An average and a corresponding standard deviation

averages_as_dict() → Dict[Measure, float]

Returns a dictionary that stores an average and a corresponding standard deviation for each measure. The averages are calculated as the arithmetic mean of all values that have been tracked for an individual measure.

Returns:

A dictionary that stores an average and a corresponding standard deviation for each measure

property measures: Set[Measure]

A set that contains all measures for which values have been tracked.

values_as_dict(index: int) → Dict[Measure, float]

Returns a dictionary that contains the value at a specific index that has been tracked for each measure.

Parameters:

index – The index of the value that should be returned for each measure

Returns:

A dictionary that stores a value for each measure

values_by_measure(measure: Measure) → ndarray

Returns an array that stores the values that have been tracked for a given measure.

Parameters:

measure – The measure

Returns:

A np.ndarray, shape (num_values_per_measure), that stores the values for the given measure