SequentialMNIST#

class torchdyno.data.datasets.seq_mnist.SequentialMNIST[source]#

Bases: MNIST

Sequential MNIST dataset.

The Sequential MNIST dataset is a variant of the MNIST dataset where the pixels of the images are permuted in a fixed way. Each image is treated pixel by pixel as a sequence, resulting in the concatenation of the rows of the image.

Methods

__init__(root[, train, transform, ...])

Sequential MNIST dataset.

Attributes

__init__(root, train=True, transform=None, target_transform=None, download=False, permute_seed=None)[source]#

Sequential MNIST dataset.

Parameters:

  • root (str, optional) – root directory of dataset.

  • train (bool, optional) – whether to load the training or test set. Defaults to True.

  • transform (Optional[Callable[..., torch.Tensor]], optional) – a function/transform that takes in an PIL image and returns a transformed version. Defaults to None.

  • target_transform (Optional[Callable[..., torch.Tensor]], optional) – a function/transform that takes in the target and transforms it. Defaults to None.

  • download (bool, optional) – whether to download the dataset. Defaults to False.

  • permute_seed (Optional[int], optional) – seed for permutation. Defaults to None.