fix n > 1 with vLLM V1 Engine

[hjh0119]

PR type

• Bug Fix
• New Feature
• Document Updates
• More Models or Datasets Support

PR information

Fix the issue where only one choice is returned when using V1 Engine and n > 1

vllm-project/vllm#7381

The same logic can be found in vllm.LLM:
https://github.com/vllm-project/vllm/blob/v0.8.5/vllm/entrypoints/llm.py#L1347-L1350

The response aggregation logic in the VLLM V1 Engine is detailed here:
https://github.com/vllm-project/vllm/blob/v0.8.5/vllm/v1/engine/parallel_sampling.py#L101-L110

Write the detail information belongs to this PR.

Experiment results

Paste your experiment result here(if needed).

[hjh0119]

test code (VLLM_USE_V1 = 1)

    from swift.llm import InferEngine, InferRequest, RequestConfig
    from swift.plugin import InferStats
    model = 'Qwen/Qwen2.5-1.5B-Instruct'
    from swift.llm import VllmEngine
    engine = VllmEngine(model, gpu_memory_utilization=0.6, use_async_engine=False, max_model_len=8192)
    req = InferRequest(messages=[{'role': 'user', 'content': 'Hello!'}])
    output = engine.infer([req], RequestConfig(n=4))