[v0.11.0-dev][Bugfix][cherry-pick]bugfix for weight load of kimi-k2

[menogrey]

What this PR does / why we need it?

This is cherry-pick from #3798

Fix kimi-k2 start bug, weight load
ERROR：#3785

Does this PR introduce any user-facing change?

How was this patch tested?

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

What this PR does / why we need it?

Does this PR introduce any user-facing change?

How was this patch tested?

[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.

[gemini-code-assist]

Code Review

This pull request correctly adds the packed module mapping for the kimi-k2 model, which is a necessary bugfix for enabling weight loading in quantized scenarios. The change is straightforward and addresses the issue. I have included one suggestion to improve the code formatting for better readability and adherence to standard Python style conventions.

[gemini-code-assist]

For better readability and adherence to Python's style guide, this list should be formatted more conventionally. The current layout, where the list value starts on a new line at the same indentation level as its key, is inconsistent with PEP 8.¹

        "experts": [
            "experts.0.gate_proj",
            "experts.0.up_proj",
            "experts.0.down_proj",
        ]

Style Guide References

Footnotes

1. PEP 8, the style guide for Python code, recommends consistent indentation for continuation lines to improve readability. The value associated with a dictionary key should either be on the same line or indented on the following lines. ↩