composer.algorithms.randaugment.randaugment

Core RandAugment code.

Functions

randaugment_image

Randomly applies a sequence of image data augmentations (Cubuk et al, 2019) to an image or batch of images.

Classes

RandAugment	Randomly applies a sequence of image data augmentations (Cubuk et al, 2019) to an image.
RandAugmentTransform	Wraps randaugment_image() in a torchvision-compatible transform.

class composer.algorithms.randaugment.randaugment.RandAugment(severity=9, depth=2, augmentation_set='all')

Bases: composer.core.algorithm.Algorithm

Randomly applies a sequence of image data augmentations (Cubuk et al, 2019) to an image.

This algorithm runs on on INIT to insert a dataset transformation. It is a no-op if this algorithm already applied itself on the State.train_dataloader.dataset.

See the Method Card for more details.

Example

from composer.algorithms import RandAugment
from composer.trainer import Trainer

randaugment_algorithm = RandAugment(
    severity=9,
    depth=2,
    augmentation_set="all"
)
trainer = Trainer(
    model=model,
    train_dataloader=train_dataloader,
    eval_dataloader=eval_dataloader,
    max_duration="1ep",
    algorithms=[randaugment_algorithm],
    optimizers=[optimizer]
)

Parameters

• severity (int, optional) – Severity of augmentation operators (between 1 to 10). M in the original paper. Default: 9.

• depth (int, optional) – Depth of augmentation chain. N in the original paper Default: 2.

• augmentation_set (str, optional) –

  Must be one of the following options:

  • "augmentations_all"

    Uses all augmentations from the paper.

  • "augmentations_corruption_safe"

    Like "augmentations_all", but excludes transforms that are part of the ImageNet-C/CIFAR10-C test sets

  • "augmentations_original"

    Like "augmentations_all", but some of the implementations are identical to the original Github repository, which contains implementation specificities for the augmentations "color", "contrast", "sharpness", and "brightness". The original implementations have an intensity sampling scheme that samples a value bounded by 0.118 at a minimum, and a maximum value of \(intensity \times 0.18 + .1\), which ranges from 0.28 (intensity = 1) to 1.9 (intensity 10). These augmentations have different effects depending on whether they are < 0 or > 0 (or < 1 or > 1). “augmentations_all” uses implementations of “color”, “contrast”, “sharpness”, and “brightness” that account for diverging effects around 0 (or 1).

  Default: "all".

class composer.algorithms.randaugment.randaugment.RandAugmentTransform(severity=9, depth=2, augmentation_set='all')

Bases: torch.nn.modules.module.Module

Wraps randaugment_image() in a torchvision-compatible transform. See RandAugment or the Method Card for more details.

Example

import torchvision.transforms as transforms
from composer.algorithms.randaugment import RandAugmentTransform

randaugment_transform = RandAugmentTransform(
    severity=9,
    depth=2,
    augmentation_set="all"
)
composed = transforms.Compose([randaugment_transform, transforms.RandomHorizontalFlip()])
transformed_image = composed(image)

Parameters

• severity (int, optional) – See RandAugment.

• depth (int, optional) – See RandAugment.

• augmentation_set (str, optional) – See RandAugment.

composer.algorithms.randaugment.randaugment.randaugment_image(img, severity=9, depth=2, augmentation_set=[<function autocontrast>, <function equalize>, <function posterize>, <function rotate>, <function solarize>, <function shear_x>, <function shear_y>, <function translate_x>, <function translate_y>, <function color>, <function contrast>, <function brightness>, <function sharpness>])

Randomly applies a sequence of image data augmentations (Cubuk et al, 2019) to an image or batch of images. See RandAugment or the Method Card for details. This function only acts on a single image (or batch of images) per call and is unlikely to be used in a training loop. Use RandAugmentTransform to use RandAugment as part of a torchvision.datasets.VisionDataset's transform.

Example

import composer.functional as cf

from composer.algorithms.utils import augmentation_sets

randaugmented_image = cf.randaugment_image(
    img=image,
    severity=9,
    depth=2,
    augmentation_set=augmentation_sets["all"]
)

Parameters

• img (Image or Tensor) – Image or batch of images to be RandAugmented.

• severity (int, optional) – See RandAugment.

• depth (int, optional) – See RandAugment.

• augmentation_set (str, optional) – See RandAugment.

Returns

PIL.Image – RandAugmented image.