dataset = tm_datasets.FC100(root=\'./data\', num_classes_per_task=5, transform=Compose([ToTensor()]) me
