Support negative indices in ndarray.__getitem__ (#107688)
In this case, we copy, but this is part of the set of divergences
described in https://github.com/Quansight-Labs/numpy_pytorch_interop/issues/73.
This does not work with dynamic shapes, but it's not clear to me what
would be the best fix
Pull Request resolved: https://github.com/pytorch/pytorch/pull/107688
Approved by: https://github.com/ezyang
ghstack dependencies: #107687
