Quantization: Add 2bit to precision constants, Fix -1 group_size handling in ModelBuilder

[jambayk]

Describe your changes

Checklist before requesting a review

• Add unit tests for this change.
• Make sure all tests can pass.
• Update documents if necessary.
• Lint and apply fixes to your code by running lintrunner -a
• Is this a user-facing change? If yes, give a description of this change to be included in the release notes.

(Optional) Issue link

[Copilot]

Pull request overview

This PR adds support for 2-bit quantization and fixes handling of per-channel quantization (group_size == -1) in the ModelBuilder pass. The changes enable proper loading and processing of quantized models that use per-channel quantization or 2-bit precision.

Changes:

• Added BITS2 = 2 to the PrecisionBits enum to support 2-bit quantization
• Fixed ModelBuilder logic to handle group_size == -1 (per-channel quantization) correctly by avoiding division by -1
• Enabled and enhanced test coverage for ModelBuilder with Olive-quantized models, including group_size == -1 test cases

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
olive/constants.py	Added BITS2 constant to PrecisionBits enum for 2-bit quantization support
olive/passes/onnx/model_builder.py	Refactored qweight tensor processing to properly handle group_size == -1 and simplified the logic by removing separate scales handling
test/passes/onnx/test_model_builder.py	Removed skip decorator and added group_size parametrization to test both normal (16) and per-channel (-1) quantization