Add aarch64 specific quantize_tensor using arm intrinsics. (#40113)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40113
Earlier version covered only armv7, aka aarch32. This diff adds aarch64 stuff
as well.
ghstack-source-id: 105990688
Test Plan: CI
Reviewed By: jerryzh168
Differential Revision: D22072779
fbshipit-source-id: c01f0b3f84394710339cf3b791832fcf68fcd4c0