tfds.dataset_builders.ViewConfig

Builder config for a view transforming another dataset.

Inherits From: BuilderConfig

tfds.dataset_builders.ViewConfig(
    *,
    name: str,
    input_dataset=None,
    ex_transformations=None,
    ds_transformations=None,
    **kwargs
)

Arguments
name	the name of this config.
input_dataset	the dataset on which this view is being applied. If None, the input dataset needs to be specified in the ViewBuilder.
ex_transformations	transformations that need to be applied to individual examples / rows. The transformations are normal Python functions. See the documentation for more information. Note that you cannot specify both example and dataset transformations.
ds_transformations	tf.data transformations that are applied on the whole tf.data.Dataset. Note that you cannot specify both example and dataset transformations.
**kwargs	additional arguments passed to the BuilderConfig such as version, release_notes, supported_versions, and description.

Attributes
name	Dataclass field
version	Dataclass field
release_notes	Dataclass field
supported_versions	Dataclass field
description	Dataclass field
tags	Dataclass field

Methods

from_dataset_info

View source

@classmethod
from_dataset_info(
    info_proto: dataset_info_pb2.DatasetInfo
) -> Optional['BuilderConfig']

Instantiates a BuilderConfig from the given proto.

Args
info_proto	DatasetInfo proto which documents the requested dataset config, including its name, version, and features.

Returns
A BuilderConfig for the requested config.

input_dataset_builder

View source

input_dataset_builder(
    data_dir: Optional[str] = None
) -> dataset_builder.DatasetBuilder

Class Variables
description	None
release_notes	None
version	None