fix(cloud-edge-llm): fix environment setup crashes and dependency typos

[rakshaak29]

Resolves #393

Description

This PR resolves critical setup failure issues in the cloud-edge-collaborative-inference-for-llm example that blocked macOS and CPU-only users from installing dependencies and running the benchmark.

Changes Made:

• requirements.txt:
  • Removed vllm (CUDA/Linux only) to make it an optional, conditionally loaded backend.
  • Corrected datamodel_code_generator to datamodel-code-generator.
  • Added missing runtime dependency retry.
• test_queryrouting.yaml:
  • Commented out the vllm backend in the default configuration so the example executes successfully out-of-the-box falling back to huggingface.

Testing Done

• Validated pip install -r requirements.txt completes successfully without CUDA.
• Ensured CloudModel successfully runs utilizing the imported retry mechanics without ImportError.

[kubeedge-bot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: rakshaak29
To complete the pull request process, please assign moorezheng after the PR has been reviewed.
You can assign the PR to them by writing /assign @moorezheng in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[gemini-code-assist]

Code Review

This pull request enhances the portability of LLM examples by replacing absolute paths with relative ones, updating the BaseModel to use a public Hugging Face model ID, and implementing dynamic device selection (CUDA/MPS/CPU). It also updates project dependencies and ignores local development artifacts. Review feedback suggests refining the MPS availability check for better compatibility, deduplicating the model identifier string, and specifying a minimum version for the torch dependency.