from openai import OpenAI

client = OpenAI(
    base_url="http://127.0.0.1:8000/v1",
    api_key="EMPTY",  # vLLM 默认通常不校验 API Key
)

response = client.chat.completions.create(
    model="qwen2.5-32b-instruct",  # 必须与 --served-model-name 一致
    messages=[
        {
            "role": "system",
            "content": "You are a helpful assistant.",
        },
        {
            "role": "user",
            "content": "简单介绍一下检索增强生成（RAG）。",
        },
    ],
    temperature=0,
    max_tokens=512,
)

print(response.choices[0].message.content)