[PyTorch] Respect existing quantizer usages in functional linear API

[timmoon10]

Description

As part of the 2.0 refactor, we exposed quantizers in the functional linear API (BasicLinear._functional_forward/BasicLinear._functional_backward). These functions internally set the quantizer usages so that the GEMMs have the data they need, but I didn't consider a case where the weight quantizer has an extra usage (i.e. so we can cache it for the backward pass). This PR modifies the functional linear API so it only sets the required usages and doesn't wipe out any user-provided usages.

Type of change

• Documentation change (change only to the documentation, either a fix or a new content)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Infra/Build change
• Code refactoring

Changes

• Respect existing quantizer usages in functional linear API

Checklist:

• I have read and followed the contributing guidelines
• The functionality is complete
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes