[Question]: qwen2.5 agent使用问题

[lonngxiang]

Has this been raised before?

• [X] I have checked the GitHub README.
• [X] I have checked the Qwen documentation and cannot find an answer there.
• [X] I have searched the issues and there is not a similar one.
• [X] I confirm that this is not a bug report, a feature request, or a badcase.

Description

运行报错：

response = client.chat.completions.create(
    model="/ai/qwen2.5/",
    messages=[{"role": "user", "content": "1+1="}],

    tools=tools,
    temperature=0.7,
    top_p=0.8,
    max_tokens=512,
    extra_body={
        "repetition_penalty": 1.05,
    },

Choice(finish_reason='stop', index=0, logprobs=None, message=ChatCompletionMessage(content='\n{{"name": "add_numbers", "arguments": {"a": 1, "b": 1}}}\n', role='assistant', function_call=None, tool_calls=[]), stop_reason=None)

[jklj077]

what is this? is it vllm? what version? how did you start it?

[lonngxiang]

what is this? is it vllm? what version? how did you start it?

vllm 0.6.1.post2

CUDA_VISIBLE_DEVICES=1 vllm serve  /ai/qwen2.5/ --host 192.16*** --port 10868 --max-model-len 2000 --trust-remote-code --api-key token-abc123 --gpu_memory_utilization 1 --trust-remote-code  --disable-frontend-multiprocessing --tensor-parallel-size 1  --enable-auto-tool-choice --tool-call-parser hermes

[jklj077]

Hi, we have identified an issue with the released chat template, which we will fix soon.

For now please try the following template:

{%- if tools %}
    {{- '<|im_start|>system\n' }}
    {%- if messages[0]['role'] == 'system' %}
        {{- messages[0]['content'] }}
    {%- else %}
        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
    {%- endif %}
    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
    {%- for tool in tools %}
        {{- "\n" }}
        {{- tool | tojson }}
    {%- endfor %}
    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
{%- else %}
    {%- if messages[0]['role'] == 'system' %}
        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
    {%- else %}
        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
    {%- endif %}
{%- endif %}
{%- for message in messages %}
    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
    {%- elif message.role == "assistant" %}
        {{- '<|im_start|>' + message.role }}
        {%- if message.content %}
            {{- '\n' + message.content }}
        {%- endif %}
        {%- for tool_call in message.tool_calls %}
            {%- if tool_call.function is defined %}
                {%- set tool_call = tool_call.function %}
            {%- endif %}
            {{- '\n<tool_call>\n{"name": "' }}
            {{- tool_call.name }}
            {{- '", "arguments": ' }}
            {{- tool_call.arguments | tojson }}
            {{- '}\n</tool_call>' }}
        {%- endfor %}
        {{- '<|im_end|>\n' }}
    {%- elif message.role == "tool" %}
        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
            {{- '<|im_start|>user' }}
        {%- endif %}
        {{- '\n<tool_response>\n' }}
        {{- message.content }}
        {{- '\n</tool_response>' }}
        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
            {{- '<|im_end|>\n' }}
        {%- endif %}
    {%- endif %}
{%- endfor %}
{%- if add_generation_prompt %}
    {{- '<|im_start|>assistant\n' }}
{%- endif %}

You can save it to, say, template.jinja and apply it as

vllm ... --enable-auto-tool-choice --tool-call-parser hermes --chat-template template.jinja 

[lonngxiang]

--chat-template template.jinja

Okay, looking forward to the repair soon.

[jklj077]

Hi, should be fixed now, please pull the model files again (updating the tokenizer_config.json is enough). Feel free to reopen if the problem persists.