APSFailureDataModule#

class torch_uncertainty.datamodules.APSFailureDataModule(root, batch_size, eval_batch_size=None, val_split=0.0, test_split=0.2, num_workers=1, pin_memory=True, persistent_workers=True, binary=True)[source]#

Tabular binary classification datamodule.

Parameters:

  • root (str | Path) – Root directory of the datasets.

  • batch_size (int) – Number of samples per batch during training.

  • eval_batch_size (int | None) – Number of samples per batch during evaluation. Defaults to batch_size.

  • val_split (float) – Share of the training samples to use as validation set. Defaults to 0.

  • test_split (float) – Share of the full dataset to hold out as test set (used when the dataset has no predefined split). Defaults to 0.2.

  • num_workers (int) – Number of data-loading subprocesses. Defaults to 1.

  • pin_memory (bool) – Whether to pin memory. Defaults to True.

  • persistent_workers (bool) – Whether to keep workers alive between epochs. Defaults to True.

  • binary (bool) – If True, returns scalar targets. Defaults to True.

dataset_class#

alias of APSFailure