mlrl.testbed.experiments.input.dataset.extension module

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

Provides classes that allow configuring the functionality to load datasets.

class mlrl.testbed.experiments.input.dataset.extension.DatasetExtension(file_type: str, *dependencies: Extension)

Bases: Extension, ABC

An abstract base class for all extensions that configure the functionality to load datasets.

abstractmethod create_sources(dataset: InputDataset, args: Namespace) → Sequence[Source]

Creates and returns one or several sources, the dataset should be loaded from.

Parameters:

• dataset – The dataset that should be loaded

• args – The command line arguments specified by the user

Returns:

A sequence that contains the sources, the dataset should be loaded from

get_supported_modes() → set[ExperimentMode]

See mlrl.testbed.extensions.extension.Extension.get_supported_modes()

class mlrl.testbed.experiments.input.dataset.extension.DatasetFileExtension(file_type: str, *dependencies: Extension)

Bases: DatasetExtension, ABC

An abstract base class for all extensions that configure the functionality to load datasets from files.

DATASET_DIRECTORY = <mlrl.util.cli.PathArgument object>

create_sources(dataset: InputDataset, args: Namespace) → Sequence[Source]

See mlrl.testbed.experiments.input.dataset.extension.DatasetExtension.create_sources()

static parse_dataset_args_from_config(config: ConfigFile) → list[ArgumentList]

Parses and returns the command line arguments for using the datasets specified in a configuration file.

Parameters:

config – The configuration file that should be parsed

Returns:

A list that contains the command line arguments that have been parsed