2024-10-26 11:34:55 +08:00
|
|
|
|
import os
|
|
|
|
|
|
import openai
|
|
|
|
|
|
|
|
|
|
|
|
# 设置OpenAI API的密钥
|
|
|
|
|
|
# openai.api_key = os.getenv("OPENAI_API_KEY")
|
2024-12-11 15:29:38 +08:00
|
|
|
|
openai.base_url = "http://127.0.0.1:8000/v1/chat/completions"
|
2024-10-26 11:34:55 +08:00
|
|
|
|
# 定义API请求的数据
|
|
|
|
|
|
data = {
|
|
|
|
|
|
"model": "chatglm3-6b",
|
|
|
|
|
|
"prompt": "Say this is a test",
|
|
|
|
|
|
"temperature": 0.5, # 控制输出结果的随机性，范围从0.0到1.0，越高越随机
|
|
|
|
|
|
"max_tokens": 75, # 控制输出文本的长度
|
|
|
|
|
|
"top_p": 1, # 一个更复杂的参数，与temperature类似但更加精细控制
|
|
|
|
|
|
"n": 1, # 要返回的最完整的文本段落数
|
|
|
|
|
|
"stream": False # 是否以流的形式返回输出
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
# 发送API请求
|
|
|
|
|
|
response = openai.Completion.create(**data)
|
|
|
|
|
|
|
|
|
|
|
|
# 打印响应结果
|
|
|
|
|
|
print(response.get("choices")[0]["text"])
|
