How to use the torchmeta.datasets.cifar100.base.CIFAR100ClassDataset function in torchmeta

.. [2] Krizhevsky A. (2009). Learning Multiple Layers of Features from Tiny
           Images. (https://www.cs.toronto.edu/~kriz/learning-features-2009-TR.pdf)
    """
    def __init__(self, root, num_classes_per_task=None, meta_train=False,
                 meta_val=False, meta_test=False, meta_split=None,
                 transform=None, target_transform=None, dataset_transform=None,
                 class_augmentations=None, download=False):
        dataset = FC100ClassDataset(root, meta_train=meta_train,
            meta_val=meta_val, meta_test=meta_test, meta_split=meta_split,
            transform=transform, class_augmentations=class_augmentations,
            download=download)
        super(FC100, self).__init__(dataset, num_classes_per_task,
            target_transform=target_transform, dataset_transform=dataset_transform)


class FC100ClassDataset(CIFAR100ClassDataset):
    subfolder = 'fc100'

    def __init__(self, root, meta_train=False, meta_val=False, meta_test=False,
                 meta_split=None, transform=None, class_augmentations=None,
                 download=False):
        super(FC100ClassDataset, self).__init__(root, meta_train=meta_train,
            meta_val=meta_val, meta_test=meta_test, meta_split=meta_split,
            transform=transform, class_augmentations=class_augmentations,
            download=download)

    def download(self):
        if self._check_integrity():
            return
        super(FC100ClassDataset, self).download()

        subfolder = os.path.join(self.root, self.subfolder)

.. [2] Krizhevsky A. (2009). Learning Multiple Layers of Features from Tiny
           Images. (https://www.cs.toronto.edu/~kriz/learning-features-2009-TR.pdf)
    """
    def __init__(self, root, num_classes_per_task=None, meta_train=False,
                 meta_val=False, meta_test=False, meta_split=None,
                 transform=None, target_transform=None, dataset_transform=None,
                 class_augmentations=None, download=False):
        dataset = CIFARFSClassDataset(root, meta_train=meta_train,
            meta_val=meta_val, meta_test=meta_test, meta_split=meta_split,
            transform=transform, class_augmentations=class_augmentations,
            download=download)
        super(CIFARFS, self).__init__(dataset, num_classes_per_task,
            target_transform=target_transform, dataset_transform=dataset_transform)


class CIFARFSClassDataset(CIFAR100ClassDataset):
    subfolder = 'cifar-fs'

    def __init__(self, root, meta_train=False, meta_val=False, meta_test=False,
                 meta_split=None, transform=None, class_augmentations=None,
                 download=False):
        super(CIFARFSClassDataset, self).__init__(root, meta_train=meta_train,
            meta_val=meta_val, meta_test=meta_test, meta_split=meta_split,
            transform=transform, class_augmentations=class_augmentations,
            download=download)

    def download(self):
        if self._check_integrity():
            return
        super(CIFARFSClassDataset, self).download()

        subfolder = os.path.join(self.root, self.subfolder)

def __init__(self, root, meta_train=False, meta_val=False, meta_test=False,
                 meta_split=None, transform=None, class_augmentations=None,
                 download=False):
        super(CIFAR100ClassDataset, self).__init__(meta_train=meta_train,
            meta_val=meta_val, meta_test=meta_test, meta_split=meta_split,
            class_augmentations=class_augmentations)

        if self.subfolder is None:
            raise ValueError()

        self.root = os.path.join(os.path.expanduser(root), self.folder)
        self.transform = transform

        self.split_filename_labels = os.path.join(self.root, self.subfolder,
            self.filename_labels.format(self.meta_split))
        self._data = None
        self._labels = None

        if download:
            self.download()