智普大模型API调用

接口

用python调用智普免费API接口的例子，写成函数，
类似于

def get_answer(prompt):
    url = 'http://34.132.32.68:8081/v1/chat/completions'
    headers = {
        'Content-Type': 'application/json',
    }
    data = {
        "model": "Qwen2-72B-int4",
        "messages": [
            {"role": "system", "content": ""},
            {"role": "user", "content": prompt}
        ],
        "temperature": 1.2,
        "max_tokens": 1000,
        "top_k": 20,
        "top_p": 0.8,
        "repetition_penalty": 1.1
    }

    response = requests.post(url, headers=headers, data=json.dumps(data))

    if response.status_code == 200:
        response_data = response.json()
        if 'choices' in response_data and len(response_data['choices']) > 0:
            return response_data['choices'][0]['message']['content']
        else:
            return "No content returned."
    else:
        return f"Request failed with status code {response.status_code}"

可用函数调用

from zhipuai import ZhipuAI


def get_answer(prompt):
    # 使用GLM-4-Flash大模型进行验证
    api_key = '51c3570a50bbaca5f9a1400'
    client = ZhipuAI(api_key=api_key)
    response = client.chat.completions.create(
        model='glm-4-flash',
        messages=[
            {'role': 'user', 'content': prompt},
        ],
        stream=True,
    )

    # 初始化输出字符串
    output = ''

    # 流式输出
    for chunk in response:
        output += chunk.choices[0].delta.content

    return output

# 使用示例
# response_text = get_glm4_flash_response('你好，我是老牛同学，请问你是谁？')
# print(response_text)