Qualcomm AI Engine Direct - Enable 4 bits BW quantization (#2506)

Summary:
- Add QNN_QUANTIZATION_ENCODING_BW... confings for qnn wrapper
- Add 4 bits quant config
- Add 4 bits quant single op tests
- Add per channel weight setting for quantizer
- Fix convert_to_linear error
- Refine quantizer

Pull Request resolved: https://github.com/pytorch/executorch/pull/2506

Reviewed By: kirklandsign

Differential Revision: D55142752

Pulled By: cccclai

fbshipit-source-id: f6ff6c753551450531e1389c50320eca42d69008