[Test]Add accuracy test for model Phi-4-mini-instruct

[MrZ20]

What this PR does / why we need it?

Does this PR introduce any user-facing change?

How was this patch tested?

• vLLM version: v0.11.0rc3
• vLLM main: vllm-project/vllm@c9461e0

[gemini-code-assist]

Code Review

This pull request adds a new accuracy test configuration for a model named Phi-4-mini-instruct. My review found a critical issue with the specified model name, as it does not seem to correspond to a publicly available model on the Hugging Face Hub. I've suggested a correction to what I believe is the intended model, microsoft/phi-3-mini-4k-instruct, and noted the related file changes that would be needed for consistency.

[gemini-code-assist]

The model name LLM-Research/Phi-4-mini-instruct appears to be incorrect. There are no public Phi-4 models from Microsoft at this time, and the LLM-Research organization does not seem to exist on the Hugging Face Hub. This will likely cause the test to fail when it attempts to download the model.

It's possible you intended to use microsoft/phi-3-mini-4k-instruct. If that's the case, please update the model name. For consistency, you should also consider renaming this YAML file to phi-3-mini-4k-instruct.yaml and updating its entry in tests/e2e/models/configs/accuracy.txt.

model_name: "microsoft/phi-3-mini-4k-instruct"

[github-actions]

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

• A PR should do only one thing, smaller PRs enable faster reviews.
• Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
• Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.