[TEST]Add initial multi modal cases for nightly test and deepseek-r1 tests

[jiangyunfan1]

What this PR does / why we need it?

This PR adds the initial multi modal model for nightly test, including 3 cases for Qwen2.5-vl-7b acc/perf test on A3, we need test them daily. It also inclues 8 cases for deepseek-r1-0528-w8a8 func, acc and perf tests

Does this PR introduce any user-facing change?

No

How was this patch tested?

by running the test

• vLLM version: v0.11.0rc3
• vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

[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 removes the --debug flag from aisbench commands and reorders test cases. My review identifies a potential robustness issue in the command generation logic in tools/aisbench.py and suggests a fix to prevent runtime errors.

[gemini-code-assist]

The two consecutive if statements for checking self.task_type can be improved. If an unknown task_type is provided, aisbench_cmd will not be defined, leading to an UnboundLocalError later. Using an if/elif/else structure will make the code more robust by explicitly handling both known cases and raising an error for any unexpected values.

        if self.task_type == "accuracy":
            aisbench_cmd = [
                'ais_bench', '--models', f'{self.request_conf}_custom',
                '--datasets', f'{dataset_conf}'
            ]
        elif self.task_type == "performance":
            aisbench_cmd = [
                'ais_bench', '--models', f'{self.request_conf}_custom',
                '--datasets', f'{dataset_conf}_custom', '--mode', 'perf'
            ]
            if self.num_prompts:
                aisbench_cmd.extend(['--num-prompts', str(self.num_prompts)])
        else:
            raise ValueError(f"Unknown task_type: {self.task_type}")