pe.data.image.imagenet module

class pe.data.image.imagenet.ImageNet(root_dir, conditional=False, split='train', res=32, batch_size=1000, num_workers=10)[source]

Bases: Data

The ImageNet dataset.

__init__(root_dir, conditional=False, split='train', res=32, batch_size=1000, num_workers=10)[source]

Constructor.

Parameters:
  • root_dir – The root directory of the dataset.

  • conditional (bool, optional) – Whether to use conditional ImageNet. Defaults to False

  • split (str, optional) – The split of the dataset, defaults to “train”

  • res (int, optional) – The resolution of the images, defaults to 32

  • batch_size (int, optional) – The batch size to load the images, defaults to 1000

  • num_workers (int, optional) – The number of workers to load the images, defaults to 10