Skip to content

semipy.datasets.SSLDataset

Warning

This section is in construction.

    class semipy.datasets.SSLDataset(dataset,
                                     weak_transform,
                                     strong_transform=None,
                                     num_classes=None):

A class based on the abstract torch.utils.data.Dataset class. It has the ability to apply two different data transformations lists (usually a strong and a weak data augmentation in SSL case). It returns a tuple of size 2 or 3, depending if the user asked a second augmentation list or not.

Parameters

  • dataset - An iterable dataset.
  • weak_transform - A first torchvision.transforms.Compose list of data transformations.
  • strong_transform - A second torchvision.transforms.Compose list of data transformations. Default: None
  • num_classes (int) - Number of classes in the 'dataset' parameter. Default: None