SEQ TINYIMAGENET R#

Classes#

class datasets.seq_tinyimagenet_r.SequentialTinyImagenet32R(args)[source]#

Bases: SequentialTinyImagenet

The Sequential TinyImagenet dataset resized to 32x32.

Parameters:

  • NAME (str) – name of the dataset.

  • SETTING (str) – setting of the dataset.

  • N_CLASSES_PER_TASK (int) – number of classes per task.

  • N_TASKS (int) – number of tasks.

  • N_CLASSES (int) – number of classes.

  • SIZE (tuple) – size of the images.

MEAN = [0.4807, 0.4485, 0.398]#
NAME: str = 'seq-tinyimg-r'#
STD = [0.2541, 0.2456, 0.2604]#
TEST_TRANSFORM = Compose(     Resize(size=32, interpolation=bilinear, max_size=None, antialias=True)     ToTensor()     Normalize(mean=[0.4807, 0.4485, 0.398], std=[0.2541, 0.2456, 0.2604]) )#
TRANSFORM = Compose(     Resize(size=32, interpolation=bilinear, max_size=None, antialias=True)     RandomCrop(size=(32, 32), padding=4)     RandomHorizontalFlip(p=0.5)     ToTensor()     Normalize(mean=[0.4807, 0.4485, 0.398], std=[0.2541, 0.2456, 0.2604]) )#
get_backbone()[source]#
get_batch_size()[source]#
get_epochs()[source]#