chatgpt接口

（1）使用curl 发送http请求

curl https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
     "model": "gpt-3.5-turbo",
     "messages": [{"role": "user", "content": "Say this is a test!"}],
     "temperature": 0.7
   }'

　请求查询gpt-3.5-turbo模型。

（2）返回结果

{
   "id":"chatcmpl-abc123",
   "object":"chat.completion",
   "created":1677858242,
   "model":"gpt-3.5-turbo-0301",
   "usage":{
      "prompt_tokens":13,
      "completion_tokens":7,
      "total_tokens":20
   },
   "choices":[
      {
         "message":{
            "role":"assistant",
            "content":"\n\nThis is a test!"
         },
         "finish_reason":"stop",
         "index":0
      }
   ]
}

　prompt_tokens:问题token数，completion_tokens:回复token数，total_tokens:总token数=问题token数+回复token数。

finish_reason:stop表示模型产生的完整回复。在上述示例中只产生了一个message,可以设置参数n来产生多个message choices.

模型

列出所有可用模型

GET https://api.openai.com/v1/models

1 2	`curl https://api.openai.com/v1/models \` `-H` `"Authorization: Bearer $OPENAI_API_KEY"`

　返回结果

{
  "data": [
    {
      "id": "model-id-0",
      "object": "model",
      "owned_by": "organization-owner",
      "permission": [...]
    },
    {
      "id": "model-id-1",
      "object": "model",
      "owned_by": "organization-owner",
      "permission": [...]
    },
    {
      "id": "model-id-2",
      "object": "model",
      "owned_by": "openai",
      "permission": [...]
    },
  ],
  "object": "list"
}

　某个模型

GET https://api.openai.com/v1/models/{model}

参数model:模型ID

1 2	`curl https://api.openai.com/v1/models/text-davinci-003 \` `-H` `"Authorization: Bearer $OPENAI_API_KEY"`

返回结果

{
  "id": "text-davinci-003",
  "object": "model",
  "owned_by": "openai",
  "permission": [...]
}

会话

创建会话(create completion)

POST https://api.openai.com/v1/completions

请求主体

请求

curl https://api.openai.com/v1/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "text-davinci-003",
    "prompt": "Say this is a test",
    "max_tokens": 7,
    "temperature": 0
  }'

　参数

{
  "model": "text-davinci-003",
  "prompt": "Say this is a test",
  "max_tokens": 7,
  "temperature": 0,
  "top_p": 1,
  "n": 1,
  "stream": false,
  "logprobs": null,
  "stop": "\n"
}

　响应

{
  "id": "cmpl-uqkvlQyYK7bGYrRHQ0eXlWi7",
  "object": "text_completion",
  "created": 1589478378,
  "model": "text-davinci-003",
  "choices": [
    {
      "text": "\n\nThis is indeed a test",
      "index": 0,
      "logprobs": null,
      "finish_reason": "length"
    }
  ],
  "usage": {
    "prompt_tokens": 5,
    "completion_tokens": 7,
    "total_tokens": 12
  }
}