Skip to content

Imagenet32

ImageNet32Dataset #

Bases: VisionDataset

Downsampled ImageNet 32x32 Dataset.

__getitem__ #

__getitem__(index)

Parameters:

Name Type Description Default
index int

Index

required

Returns: tuple: (image, target) where target is index of the target class.

ImageNet32DataModule #

Bases: VisionDataModule

TODO: Add a val_split argument, that supports a value of 0.

prepare_data #

prepare_data() -> None

Saves files to data_dir.

default_transforms #

default_transforms() -> Callable

Default transform for the dataset.

train_dataloader #

train_dataloader() -> DataLoader

The train dataloader.

val_dataloader #

val_dataloader() -> DataLoader

The val dataloader.

test_dataloader #

test_dataloader() -> DataLoader

The test dataloader.

get_train_val_indices #

get_train_val_indices(
    dataset_labels: Sequence[int] | ndarray,
    nb_imgs_in_val: int,
    split_seed: int,
) -> tuple[list[int], list[int]]

Keeps the first nb_imgs_in_val images of each class in the validation set.