DOTA2GamesDataModule#

class torch_uncertainty.datamodules.DOTA2GamesDataModule(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]#

The Dota2 Games UCI classification datamodule.

Parameters:

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

  • batch_size (int) – The batch size for training and testing.

  • eval_batch_size (int | None) – Number of samples per batch during evaluation (val and test). Set to batch_size if None. Defaults to None.

  • val_split (float) – Share of validation samples among the non-test samples. Defaults to 0.

  • test_split (float) – Share of test samples. Defaults to 0.2.

  • num_workers (int) – How many subprocesses to use for data loading. Defaults to 1.

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

  • persistent_workers (bool) – Whether to use persistent workers. Defaults to True.

  • binary (bool) – Whether to use binary classification. Defaults to True.