incerto.data.MNIST

class incerto.data.MNIST(root='./data', val_split=0.1, seed=42, normalize=True)

Bases: VisionDataset

MNIST dataset with standardized splits.

28x28 grayscale images of handwritten digits (0-9). 60,000 training + 10,000 test images.

Parameters:

• root (str | Path)

• val_split (float)

• seed (int)

• normalize (bool)

__init__(root='./data', val_split=0.1, seed=42, normalize=True)

Initialize vision dataset.

Parameters:

• root (str | Path) – Root directory for dataset storage

• val_split (float) – Fraction of training data to use for validation

• seed (int) – Random seed for reproducible splits

• normalize (bool)

Methods

__init__([root, val_split, seed, normalize])	Initialize vision dataset.
get_datasets()	Get train, validation, and test datasets.
get_transforms([train])	Get data transformations.

__init__(root='./data', val_split=0.1, seed=42, normalize=True)

Initialize vision dataset.

Parameters:

• root (str | Path) – Root directory for dataset storage

• val_split (float) – Fraction of training data to use for validation

• seed (int) – Random seed for reproducible splits

• normalize (bool)

get_transforms(train=False)

Get data transformations.

Parameters:

train (bool)

Return type:

Compose

get_datasets()

Get train, validation, and test datasets.

Return type:

Tuple[Dataset, Dataset, Dataset]

Returns:

Tuple of (train_dataset, val_dataset, test_dataset)