Update Quantization Scheme Standards for better readability

[markurtz]

Summary

Enhances the quantization module by expanding the quantization schemes, inline docs, and default values to improve readability and usability.

Details

• Added new quantization schemes for 4-bit weights and 8-bit activations (W4A8) and improved documentation for existing schemes.
• Updated the QuantizationArgs to include comprehensive default values.
• Added docs throughout to clarify the code

Test Plan

• Manual review
• Current automated tests

[dbogunowicz]

It's a copy of: #105 isnt it @markurtz ?

[markurtz]

It's a copy of: #105 isnt it @markurtz ?

#105 landed, but didn't include the fixes I wanted to get in with it. I've rebased on current state of main so they no longer contain duplicate code