old_mnist_360#

Classes

class datasets.deprecated.old_mnist_360.MNIST360(args)[source]#

Bases: GCLDataset

MNIST-360 general continual dataset.

LENGTH = 54051#
NAME: str = 'old-mnist-360'#
N_CLASSES: int = 9#
SETTING: str = 'general-continual'#
static get_backbone()[source]#
Return type:

Module

static get_batch_size()[source]#
Return type:

int

static get_denormalization_transform()[source]#
static get_loss()[source]#
Return type:

cross_entropy

static get_minibatch_size()[source]#
Return type:

int

get_test_data()[source]#

Ensembles the next examples of the current class in a batch. :return: the batch of examples along with its label.

Return type:

Tuple[Tensor, Tensor]

get_train_data()[source]#

Ensembles the next examples of the current classes in a single batch. :return: the augmented and not aumented version of the examples of the

current batch, along with their labels.

Return type:

Tuple[Tensor, Tensor, Tensor]

static get_transform()[source]#
init_test_loaders()[source]#

Initializes the test loader.

init_train_loaders()[source]#

Initializes the test loader.

train_next_class()[source]#

Changes the couple of current training classes.