mlrl.testbed.experiments.recipe module

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

Provides classes for setting up experiments.

class mlrl.testbed.experiments.recipe.Recipe

Bases: ABC

An abstract base class for all classes that provide access to the ingredients that are needed by different operational modes for setting up experiments.

abstractmethod create_dataset_splitter(args: Namespace, load_dataset: bool = True) → DatasetSplitter

Creates and returns the DatasetSplitter to be used in experiments.

Parameters:

  • args – The command line arguments specified by the user

  • load_dataset – True, if the dataset should be loaded by the DatasetSplitter, False otherwise

Returns:

The DatasetSplitter that has been created

abstractmethod create_experiment_builder(experiment_mode: ExperimentMode, args: Namespace, command: Command, load_dataset: bool = True) → Builder

Creates and returns the Experiment.Builder to be used for configuring experiments.

Parameters:

  • experiment_mode – The mode of operation

  • args – The command line arguments specified by the user

  • command – The command that has been used to start the experiment

  • load_dataset – True, if the dataset should be loaded by the experiment, False otherwise

Returns:

The Experiment.Builder that has been created

abstractmethod create_problem_domain(mode: ExperimentMode, args: Namespace) → ProblemDomain

Creates and returns the ProblemDomain to be used in experiments.

Parameters:

  • mode – The mode of operation

  • args – The command line arguments specified by the user

Returns:

The ProblemDomain that has been created