nip.image_classification.data.TorchVisionDatasetWrapper

class nip.image_classification.data.TorchVisionDatasetWrapper(root: str | Path, train: bool = True, transform: callable | None = None, target_transform: callable | None = None, download: bool = False)

A wrapper for TorchVision datasets, implementing a common interface.

Derived classes should defined the class attributes below.

Class attributes

data_classtype[VisionDataset]

The TorchVision dataset class.

num_channelsint

The number of channels of the images in the dataset.

widthint

The width of the images in the dataset.

heightint

The height of the images in the dataset.

selected_classestuple[int, int]

When selecting two classes from the original dataset, the default two to select.

binarification_seedint

The seed used when doing a randomised binarification.

Methods Summary

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

Attributes

binarification_seed
selected_classes
num_channels
width
height

Methods

__init__(root: str | Path, train: bool = True, transform: callable | None = None, target_transform: callable | None = None, download: bool = False)