openai functioncall 에러 메세지

소모임 > 인공지능 Posted 2023.11.10 by 티쳐✨

899

Error code: 400 - {'error': {'message': "Invalid parameter: messages with role 'tool' must be a response to a preceeding message with 'tool_calls'.", 'type': 'invalid_request_error', 'param': 'messages.[1].role', 'code': None}}

이번 업데이트된 버전으로 코드를 수정하고 있습니다.

그러던 중 functioncall 기능을 이용한 코드를 수정하였는데, 위와 같은 오류가 발생했습니다.

코드의 구조는 openai에서 제시하고 있는 기본 코드 구조인 다음의 형태를 그대로 따랐지만....

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
from openai import OpenAI
import json
 
client = OpenAI()
 
# Example dummy function hard coded to return the same weather
# In production, this could be your backend API or an external API
def get_current_weather(location, unit="fahrenheit"):
    """Get the current weather in a given location"""
    if "tokyo" in location.lower():
        return json.dumps({"location": "Tokyo", "temperature": "10", "unit": "celsius"})
    elif "san francisco" in location.lower():
        return json.dumps({"location": "San Francisco", "temperature": "72", "unit": "fahrenheit"})
    elif "paris" in location.lower():
        return json.dumps({"location": "Paris", "temperature": "22", "unit": "celsius"})
    else:
        return json.dumps({"location": location, "temperature": "unknown"})
 
def run_conversation():
    # Step 1: send the conversation and available functions to the model
    messages = [{"role": "user", "content": "What's the weather like in San Francisco, Tokyo, and Paris?"}]
    tools = [
        {
            "type": "function",
            "function": {
                "name": "get_current_weather",
                "description": "Get the current weather in a given location",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "location": {
                            "type": "string",
                            "description": "The city and state, e.g. San Francisco, CA",
                        },
                        "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
                    },
                    "required": ["location"],
                },
            },
        }
    ]
    response = client.chat.completions.create(
        model="gpt-3.5-turbo-1106",
        messages=messages,
        tools=tools,
        tool_choice="auto",  # auto is default, but we'll be explicit
    )
    response_message = response.choices[0].message
    tool_calls = response_message.tool_calls
    # Step 2: check if the model wanted to call a function
    if tool_calls:
        # Step 3: call the function
        # Note: the JSON response may not always be valid; be sure to handle errors
        available_functions = {
            "get_current_weather": get_current_weather,
        }  # only one function in this example, but you can have multiple
        messages.append(response_message)  # extend conversation with assistant's reply
        # Step 4: send the info for each function call and function response to the model
        for tool_call in tool_calls:
            function_name = tool_call.function.name
            function_to_call = available_functions[function_name]
            function_args = json.loads(tool_call.function.arguments)
            function_response = function_to_call(
                location=function_args.get("location"),
                unit=function_args.get("unit"),
            )
            messages.append(
                {
                    "tool_call_id": tool_call.id,
                    "role": "tool",
                    "name": function_name,
                    "content": function_response,
                }
            )  # extend conversation with function response
        second_response = client.chat.completions.create(
            model="gpt-3.5-turbo-1106",
            messages=messages,
        )  # get a new response from the model where it can see the function response
        return second_response
print(run_conversation())
Colored by Color Scripter
cs

오류가 발생했습니다.

그래서 그냥

                    "tool_call_id": tool_call.id,

                    "role": "tool",

                    "name": function_name,

                    "content": function_response,

이부분을 지우고

role : assistant 로 넣고 content를 function_respose로 붙여서 임시 방편으로 해결했습니다.

혹시 정확한 해결방법을 아시는 분은 알려주세요!!

#chatgpt, #에러

티쳐✨ 전라남도

소개

::: Foreducator.com을 운영하는 수학교사 ::: '수학하는 즐거움 시리즈', '수업의 과정' 저자

로그인 후 댓글을 작성하실 수 있습니다.

Activity Logs

Database Backup Process Completed!

Shopping Cart

What’s in Startup Plan?

What’s in Startup Plan?

What’s in Startup Plan?

What’s in Startup Plan?

Details

Frameworks

Database

Billing

Completed

Drop campaign files here or click to upload.

Project Type

Project Settings

Drop files here or click to upload.

Budget

Build a Team

Set First Target

Upload Files

Drop files here or click to upload.

Project Created!

Users

Customers

Projects

No result found

Advanced Search

Search Preferences

Notifications

Get Pro Access

Activity Logs

Database Backup Process Completed!

Shopping Cart

Drop files here or click to upload.

Drop files here or click to upload.