Move GroupedBatchSampler into torchvision

🚀 Feature

Move the GroupedBatchSampler into the core library

Motivation

Grouping minibatch elements is often a useful feature for vision tasks, especially in object detection and segmentation problems where it is commonplace to use images with different shapes and aspect ratios. Torchvision already has an implementation of a sampler that can do this, however it isn't part of the library itself, only in the references. As such users either have to copy-paste it into their own code or import detectron2 to utilise it.

Pitch

Probably just copy the entire group_by_aspect_ratio.py into torchvision/datasets/samplers, perhaps split it out if necessary. I think this would all be useful for people working with detection or segmentation tasks that don't want to bring in all of d2.

Start with references/detection/group_by_aspect_ratio.py, especially the GroupedBatchSampler implementation, and compare it with the proposed torchvision/datasets/samplers location. Determine how the sampler should be exposed as part of the core library and preserve the referenced behavior; done means users can use it from torchvision without copying the reference code or importing detectron2.