Add support for tensor-wise mixed precision in Quark onnx quantization

[gengxinwu]

Describe your changes

Add support for tensor-wise mixed precision in Quark onnx quantization.

It supports specifying quantization data types for specific tensors, how to configure it please refer to the newly added testcase.

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

[jambayk]

@gengxinwu please agree to the CLA following the bot instruction above. thanks!