Configuration

These ray.train classes configure a training run: how many workers to use, where to store results, and how to checkpoint, validate, and recover from failures.

CheckpointConfig	Configuration for checkpointing.
DataConfig	Class responsible for configuring Train dataset preprocessing.
FailureConfig	Configuration related to failure handling of each training run.
LoggingConfig	Configuration for Ray Train's logging behavior.
RunConfig	Runtime configuration for training runs.
ScalingConfig	Configuration for scaling training.
ValidationConfig	Configuration for validation, passed to the trainer.