Use torchvision's transforms v2

[dxoigmn]

Torchvision's v2 transforms support for transform images and bounding boxes similar to how we implement via ExTransform: https://github.com/IntelLabs/MART/blob/20d20789548c232e8d9a13079c2cef222fb9889b/mart/transforms/extended.py#L35

I think we should just standardize on Torchvision's implementation.

https://pytorch.org/vision/main/auto_examples/plot_transforms_v2.html

[dxoigmn]

See also: https://pytorch.org/vision/stable/auto_examples/plot_transforms_v2_e2e.html#sphx-glr-auto-examples-plot-transforms-v2-e2e-py