mindformers.models.mae.ViTMAEImageProcessor

class mindformers.models.mae.ViTMAEImageProcessor(size=224, patch_size=16, mask_ratio=0.75)

ViTMAEImageProcessor.

Args:

image_resolution (int): the target size.

preprocess(images, **kwargs)

Preprocess required by base processor.

Args:

images (tensor, PIL.Image, numpy.array, list): a batch of images.

Return:

A 4-rank tensor for a batch of images.