DataModule(name: str = 'bookcorpus', split: str = 'train', batch_size: int = 4, num_workers: int = 16, pin_memory: bool = True)
Bases: LightningDataModule
Initializes a DataLoader object for "bookcorpus". Support for more datasets coming soon.
Examples:
setup
Initializes a huggingface dataset: bookcorpus.
train_dataloader
train_dataloader() -> DataLoader
Creates instance of DataLoader
.
Returns:
-
DataLoader
–
A DataLoader for a specified dataset.