update awq backend infos, refine UT

[WeiweiZhang1]

Description

Please briefly describe your main changes, the motivation.

Type of Change

• Bug fix
• New feature
• Documentation update
• Performance improvement
• Code refactoring
• Other (please specify):

Related Issues

Fixes or relates to #

Checklist Before Submitting

• My code has been tested locally.
• Documentation has been updated as needed.
• New or updated tests are included where applicable.

[Copilot]

Pull request overview

Updates AWQ backend documentation (including gptqmodel-based AWQ variants) and refines a CUDA integration test to explicitly release SGLang resources.

Changes:

• Add explicit llm.shutdown() in SGLang CUDA integration test.
• Expand AWQ backend matrix docs to include gptqmodel AWQ variants and add a Transformers version constraint for auto-awq.
• Clean up/align backend table formatting in the English step-by-step doc.

Reviewed changes

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

File	Description
test/test_cuda/integrations/test_sglang.py	Ensures SGLang inference test shuts down the LLM instance after generation.
docs/step_by_step_CN.md	Updates backend capability table with AWQ-related backend entries and requirements.
docs/step_by_step.md	Re-formats backend selection docs/table and adds AWQ backend requirement details.