AutoAugment

class mmcls.datasets.transforms.AutoAugment(policies, hparams={'pad_val': 128})

Auto augmentation.

This data augmentation is proposed in AutoAugment: Learning Augmentation Policies from Data.

参数

• policies (str | list[list[dict]]) – The policies of auto augmentation. If string, use preset policies collection like “imagenet”. If list, Each item is a sub policies, composed by several augmentation policy dicts. When AutoAugment is called, a random sub policies in policies will be selected to augment images.

• hparams (dict) – Configs of hyperparameters. Hyperparameters will be used in policies that require these arguments if these arguments are not set in policy dicts. Defaults to dict(pad_val=128).