Refactor arm quantization (#3616)

Summary:
* Move quantization files to own folder
* Breakout annotators into one file/ annotator in new folder quantizer/quantization_annotation
* Make QuantizationConfig not optional in annotators There is a None-check before calling them.
* Move all Quantization config logic into one class
* Remove unsupported qat functionality
* Remove unsupported dynamic annotation
* Remove unused code
* Add docstrings

Change-Id: Ie802780bd0eb7f946985e09de13c822645a78e21

Pull Request resolved: https://github.com/pytorch/executorch/pull/3616

Reviewed By: manuelcandales

Differential Revision: D57618668

Pulled By: digantdesai

fbshipit-source-id: ec56bcc0862b7e6d69ffb0d3883c35ba5d2087f5