ray.train.DataConfig

class ray.train.DataConfig(datasets_to_split: Union[Literal['all'], List[str]] = 'all', execution_options: Optional[ray.data._internal.execution.interfaces.execution_options.ExecutionOptions] = None)

Bases: object

Class responsible for configuring Train dataset preprocessing.

For advanced use cases, this class can be subclassed and the configure() method overriden for custom data preprocessing.

PublicAPI: This API is stable across Ray releases.

Methods

__init__([datasets_to_split, execution_options])	Construct a DataConfig.
configure(datasets, world_size, ...)	Configure how Train datasets should be assigned to workers.
default_ingest_options()	The default Ray Data options used for data ingest.