qwen2-vl-7b-instruct 以VLLM形式启动推理引擎失败“ assert "factor" in rope_scaling”

[wyclike]

Describe the bug
What the bug is, and how to reproduce, better with screenshots(描述bug以及复现过程，最好有截图)
命令：CUDA_VISIBLE_DEVICES=0 swift infer --model_type qwen2-vl-7b-instruct --infer_backend vllm --max_model_len 4096

[INFO:swift] Loading the model using model_dir: /vepfs/fs_users/weiyuancheng-test/ckpt/hub/qwen/Qwen2-VL-7B-Instruct
Traceback (most recent call last):
File "/vepfs/fs_users/weiyuancheng-test/swift/swift/cli/infer.py", line 5, in
infer_main()
File "/vepfs/fs_users/weiyuancheng-test/swift/swift/utils/run_utils.py", line 32, in x_main
result = llm_x(args, **kwargs)
File "/vepfs/fs_users/weiyuancheng-test/swift/swift/llm/infer.py", line 282, in llm_infer
llm_engine, template = prepare_vllm_engine_template(args)
File "/vepfs/fs_users/weiyuancheng-test/swift/swift/llm/utils/vllm_utils.py", line 583, in prepare_vllm_engine_template
llm_engine = get_vllm_engine(
File "/vepfs/fs_users/weiyuancheng-test/swift/swift/llm/utils/vllm_utils.py", line 103, in get_vllm_engine
llm_engine = llm_engine_cls.from_engine_args(engine_args)
File "/vepfs/fs_users/weiyuancheng-test/envs/main/lib/python3.10/site-packages/vllm/engine/llm_engine.py", line 438, in from_engine_args
engine_config = engine_args.create_engine_config()
File "/vepfs/fs_users/weiyuancheng-test/envs/main/lib/python3.10/site-packages/vllm/engine/arg_utils.py", line 699, in create_engine_config
model_config = ModelConfig(
File "/vepfs/fs_users/weiyuancheng-test/envs/main/lib/python3.10/site-packages/vllm/config.py", line 167, in init
self.max_model_len = _get_and_verify_max_len(
File "/vepfs/fs_users/weiyuancheng-test/envs/main/lib/python3.10/site-packages/vllm/config.py", line 1506, in _get_and_verify_max_len
assert "factor" in rope_scaling
AssertionError

Your hardware and system info
Write your system info like CUDA version/system/GPU/torch version here(在这里给出硬件信息和系统信息，如CUDA版本，系统，GPU型号和torch版本等)
4090
torch：2.3.1

[zhangfan-algo]

＋1

[wangaocheng]

遇到相同的问题。

[Jintao-Huang]

QwenLM/Qwen2-VL#96