ci: unify the model launch method of nightly ci

[mickqian]

Motivation

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.

[mickqian]

Guide for adding models to nightly-ci

• for text models: extend global model lists variables prefixed with DEFAULT_MODEL_NAME_FOR_NIGHTLY_EVAL_ in test_utils.py, or add more model lists
• for vlms: extend global variables containing ModelLaunchSettings in test_nightly_vlms_.*.py, see here

[JustinTong0323]

Maybe use

VLMModelDeploySetup = partial(ModelLaunchSettings, extra_args=VLM_EXTRA_ARGS)
VLMModelLaunchSettings("zai-org/GLM-4.1V-9B-Thinking"): ModelEvalMetrics(0.280, 30.4),

to avoid code duplication

[mickqian]

fixed