【新增功能需求】 Internvl2模型+VLLM 后端实现 异步客户端请求的Video 推理功能

[PancakeAwesome]

Describe the feature Please describe the feature requested here(请在这里描述需求) Internvl2模型+VLLM 后端实现 异步客户端请求的Video 推理功能 Paste any useful information Paste any useful information, including papers, github links, etc.(请在这里描述其他有用的信息，比如相关的论文地址，github链接等) 现在的 vllm 2.4.0dev 分支运行以下 异步客户端请求逻辑会出错。 swift deploy --infer_backend vllm --max_model_len 2048 --host 127.0.0.1 --port 9122 --model_type internvl2-8b

from swift.llm import inference_client_async
question = '描述这个视频'
query = '<video>' + question
videos = ['./1.mp4']
tasks = [inference_client_async(model_type, query, videos=videos, request_config=request_config, port='9122') for _ in range(1)]
resp_list = asyncio.run(_batch_run(tasks))

>>>>>>>>>>>>>>>ERROR
ContentTypeError: 0, message='Attempt to decode JSON with unexpected mimetype: text/plain; charset=utf-8', url=URL('http://127.0.0.1:9122/v1/chat/completions')

Additional context Add any other context or information here(其他信息可以写在这里) ms-swift 2.4.0dev vllm 0.5.4

[PancakeAwesome]

@Jintao-Huang Looking forward to your reply, Thank u!