MNIST 360

Classes

class datasets.mnist_360.MNIST360(args, is_train=False)

Bases: Dataset

A custom dataset class for MNIST360 that provides training and testing data with incremental rotation for each class.

Parameters:

• args (object) – An object containing the arguments for the dataset.

• is_train (bool) – A flag indicating whether the dataset is for training or testing.

N_CLASSES

The number of classes in the dataset.

Type:

int

dataset

A list of data loaders for each class.

Type:

list

remaining_training_items

A list of the remaining training items for each class.

Type:

list

num_rounds

The number of rounds for each class.

Type:

int

args

An object containing the arguments for the dataset.

Type:

object

is_train

A flag indicating whether the dataset is for training or testing.

Type:

bool

is_over

A flag indicating whether the dataset is completed.

Type:

bool

completed_rounds

The number of completed rounds.

Type:

int

test_class

The current test class index.

Type:

int

test_iteration

The current test iteration index.

Type:

int

train_classes

A list of the current training classes.

Type:

list

active_train_loaders

A list of the active training data loaders.

Type:

list

current_items

The current number of items in the dataset.

Type:

int

N_CLASSES = 9

get_test_data()

Ensembles the next examples of the current class in a batch.

Returns:

The batch of examples. Tensor: The labels of the examples.

Return type:

Tensor

get_train_data()

Ensembles the next examples of the current classes in a single batch.

Returns:

The batch of examples.

Tensor: The labels of the examples.

Tensor: The batch of examples without augmentation.

Return type:

Tensor

init_test_loaders()

Initializes the test loader.

init_train_loaders()

Initializes the train loader.

reinit()

train_next_class()

Changes the couple of current training classes.

class datasets.mnist_360.SequentialMNIST360(args)

Bases: GCLDataset

A dataset class for the MNIST-360 dataset in the context of general-continual learning.

NAME

The name of the dataset.

Type:

str

SETTING

The setting of the dataset.

Type:

str

N_CLASSES

The number of classes in the dataset.

Type:

int

TRANSFORM

The transformation to apply to the data.

Type:

torch.nn.Module

SIZE

The size of the input images.

Type:

tuple

args

An object containing the arguments for the dataset.

Type:

Namespace

NAME: str = 'mnist-360'

N_CLASSES: int = 9

SETTING: str = 'general-continual'

SIZE: Tuple[int] = (28, 28)

TRANSFORM = Identity()

get_backbone()

Return type:

Module

get_batch_size()

Return type:

int

get_data_loaders()

Get the data loaders for the MNIST360 dataset, add them to the current object and return them.

Returns:

DataLoader for the training dataset.

test_loader (torch.utils.data.DataLoader): DataLoader for the test dataset.

Return type:

train_loader (torch.utils.data.DataLoader)

static get_denormalization_transform()

get_epochs()

static get_loss()

Return type:

Callable

static get_normalization_transform()

static get_transform()

Functions

datasets.mnist_360.custom_collate_unbatch(batch)

Custom collate function to unbatch a batch of data.

Parameters:

batch (list) – A list of tensors representing a batch of data.

Returns:

A list of tensors, where each tensor is unbatched from the input batch.

Return type:

list

datasets.mnist_360.custom_collate_unbatch(batch)

Custom collate function to unbatch a batch of data.

Parameters:

batch (list) – A list of tensors representing a batch of data.

Returns:

A list of tensors, where each tensor is unbatched from the input batch.

Return type:

list