Setting more than 2 elements to `scale` argument of `RandomResizedCrop()` works

[hyperkai]

🐛 Describe the bug

Setting more than 2 elements to scale argument of RandomResizedCrop() works as shown below. *scale argument should accept 2 elements:

from torchvision.transforms.v2 import RandomResizedCrop

rrc = RandomResizedCrop(size=100, scale=[0.1, 0.2, 0.3, 0.4, 0.5])

rrc
# RandomResizedCrop(size=(100, 100),
#                   scale=[0.1, 0.2, 0.3, 0.4, 0.5],
#                   ratio=(0.75, 1.3333333333333333),
#                   interpolation=InterpolationMode.BILINEAR,
#                   antialias=True)

In addition, setting 0 or 1 element to scale argument of RandomResizedCrop() doesn't work as shown below:

from torchvision.transforms.v2 import RandomResizedCrop

rrc = RandomResizedCrop(size=100, scale=[])

rrc # Error

rrc = RandomResizedCrop(size=100, scale=[0.1])

rrc # Error

IndexError: list index out of range

Versions

import torchvision

torchvision.__version__ # '0.20.1'

[NicolasHug]

Feel free to submit a PR @hyperkai , thanks