Module: tfds.core

API to define datasets.

Classes

class BeamBasedBuilder: Beam based Builder.

class BeamMetadataDict: A tfds.core.Metadata object supporting Beam-generated datasets.

class BuilderConfig: Base class for DatasetBuilder data configuration.

class DatasetBuilder: Abstract base class for all datasets.

class DatasetInfo: Information about a dataset.

class Experiment: Experiments which can be enabled/disabled on a per version basis.

class GeneratorBasedBuilder: Base class for datasets with data generation based on file adapter.

class Metadata: Abstract base class for DatasetInfo metadata container.

class MetadataDict: A tfds.core.Metadata object that acts as a dict.

class ReadInstruction: Reading instruction for a dataset.

class ReadOnlyPath: Protocol for read-only methods of pathlib.Path-like API.

class ReadWritePath: Protocol for pathlib.Path-like API.

class SplitDict: Split info object.

class SplitGenerator: Defines the split information for the generator.

class SplitInfo: Wraps proto.SplitInfo with an additional property.

class Version: Dataset version MAJOR.MINOR.PATCH.

class lazy_imports: Lazy importer for heavy dependencies.

Functions

add_data_dir(...): Registers a new default data_dir to search for datasets.

as_path(...): Create a generic pathlib.Path-like abstraction.

benchmark(...): Benchmarks a tf.data.Dataset.

builder_from_directory(...): Loads a tfds.core.DatasetBuilder from the given generated dataset path.

gcs_path(...): Returns the GCS URI path.

tfds_path(...): Path to tensorflow_datasets/ root dir.