add support for apply probability to CutMix and MixUp

[pmeier]

No description provided.

[pmeier]

Default for other transforms is p=0.5:

vision/torchvision/prototype/transforms/_augment.py

Lines 15 to 18 in 2a0eea8

	class RandomErasing(_RandomApplyTransform):
	def __init__(
	self,
	p: float = 0.5,

vision/references/classification/transforms.py

Line 22 in 2a0eea8

def __init__(self, num_classes: int, p: float = 0.5, alpha: float = 1.0, inplace: bool = False) -> None:

In the classification reference we hardcode p=1.0 for CutMix and MixUp:

vision/references/classification/train.py

Lines 204 to 207 in 2a0eea8

	if args.mixup_alpha > 0.0:
	mixup_transforms.append(transforms.RandomMixup(num_classes, p=1.0, alpha=args.mixup_alpha))
	if args.cutmix_alpha > 0.0:
	mixup_transforms.append(transforms.RandomCutmix(num_classes, p=1.0, alpha=args.cutmix_alpha))

I would keep p=0.5 here and set p=1.0 in the references as before. No strong opinion though.

[datumbox]

LGTM, thanks!