[Bug]: qwen2.5 tools输出格式问题

[kwjlhh]

Model Series

Qwen2.5

What are the models used?

Qwen2.5-72B-Instruct

What is the scenario where the problem happened?

deployment with vllm， tool calling with vllm

Is this a known issue?

• [X] I have followed the GitHub README.
• [X] I have checked the Qwen documentation and cannot find an answer there.
• [X] I have checked the documentation of the related framework and cannot find useful information.
• [X] I have searched the issues and there is not a similar one.

Information about environment

Ubuntu 22.04 Python 3.10.14 vllm 0.6.2

Log output

log is fine

Description

有概率出现如下结构： "message": { "role": "assistant", "content": "\n{\"name\": \"text_classification\", \"arguments\": {\"classifications\": [{\"classification\": \"test\", \"score\": 0.9563}, {\"classification\": \test2\", \"score\": 0.0321}, {\"classification\": \"test3\", \"score\": 0.0102}]}\n", "tool_calls": [] }, 模型代码文件都是9.25号拉的，应该是晚于之前issues所提到的更新 tokenizer_config.json的时间点

[jklj077]

please provide the input and the sampling parameters so that we can try reproducing the issue.