**Describe the bug** 当我使用本项目评测我本地启动的模型服务时,跑一会就会发生报错 <img width="1250" height="570" alt="Image" src="https://github.com/user-attachments/assets/d8c83cf8-e95b-4d27-879a-9ecd597c2b2f" /> **start command** python run.py benchmark --target "http://10.16.2.3:11092" --model Qwen2_5-VL-7B-Instruct --processor /Users/huizhang/Develop/ai/model/Qwen2.5-VL-7B-Instruct --rate-type sweep --max-seconds 60 --data "prompt_tokens=512,output_tokens=10" **Environment** Include all relevant environment information: 1. OS: macOS 12.7.3 2. Python version : 3.10.15 帮忙确定一下是什么环节的问题