AttributeError: module 'openai' has no attribute 'ChatCompletion'

[monk1337]

Describe the bug

The feature was working few hours ago, now it's giving error.

To Reproduce

import openai

openai.api_key = 'key'

openai.ChatCompletion.create(
  model="gpt-3.5-turbo",
  messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Who won the world series in 2020?"},
        {"role": "assistant", "content": "The Los Angeles Dodgers won the World Series in 2020."},
        {"role": "user", "content": "Where was it played?"}
    ]
)

Code snippets

No response

OS

macOS

Python version

Python v3.7.1

Library version

openai v0.26.5

no function as ChatCompletion the defined methods

import openai
print(openai.__dir__())

['name', 'doc', 'package', 'loader', 'spec', 'path', 'file', 'cached', 'builtins', 'annotations', 'os', 'Optional', 'error', 'util', 'version', 'openai_response', 'api_requestor', 'openai_object', 'api_resources', 'Answer', 'Classification', 'Completion', 'Customer', 'Edit', 'Deployment', 'Embedding', 'Engine', 'ErrorObject', 'File', 'FineTune', 'Image', 'Model', 'Moderation', 'Search', 'APIError', 'InvalidRequestError', 'OpenAIError', 'api_key', 'api_key_path', 'organization', 'api_base', 'api_type', 'api_version', 'verify_ssl_certs', 'proxy', 'app_info', 'enable_telemetry', 'ca_bundle_path', 'debug', 'log', 'all']

[logankilpatrick]

Please bump your openai v0.26.5 to v0.27.0.

[elias-py]

You should have the new version v.27.0

[onesecurity]

running the command:

pip install --upgrade openai
Alternatively, to install it manually, do the following:

1. Download openai-0.27.0-py3-none-any.whl from https://pypi.org/project/openai/#files
2. Run pip install openai-0.27.0-py3-none-any.whl

[captain-miao]

openai-0.27.0

2023-03-02 11:16:49,016 - openai - INFO - error_code=None error_message='Invalid URL (POST /v1/engines/gpt-3.5-turbo/chat/completions)' error_param=None error_type=invalid_request_error message='OpenAI API error received' stream_error=False

[captain-miao]

openai-0.27.0

2023-03-02 11:16:49,016 - openai - INFO - error_code=None error_message='Invalid URL (POST /v1/engines/gpt-3.5-turbo/chat/completions)' error_param=None error_type=invalid_request_error message='OpenAI API error received' stream_error=False

It's already OK, caused by parameter error.

        response =  openai.ChatCompletion.create(
            model="gpt-3.5-turbo-0301",
            messages=[{"role": "user", "content": update.message.text}],
            max_tokens=1024,
            n=1,
            temperature=0.5,
        )
        update.message.reply_text(response["choices"][0]["message"]["content"])

[hanswang73]

running the command:

pip install --upgrade openai Alternatively, to install it manually, do the following:

1. Download openai-0.27.0-py3-none-any.whl from https://pypi.org/project/openai/#files
2. Run pip install openai-0.27.0-py3-none-any.whl

I do it, but the problem do not disappear.

[mizumot]

Restarting Python itself may solve the problem. It worked for me.

[emilycowe]

I'm getting the same issue, despite using openai-0.27.0. Does anyone have any ideas of what else I can try?

[monk1337]

I solved the issue by upgrading the openai v27.0 by

pip install --upgrade openai
Example Usage:

ChatGPT API has been added in Promptify; you can use it like this.

from promptify import OpenAI
from promptify import Prompter

sentence     =  """The patient is a 93-year-old female with a medical                   
                history of chronic right hip pain, osteoporosis,                    
                hypertension, depression, and chronic atrial                        
                fibrillation admitted for evaluation and management                
                of severe nausea and vomiting and urinary tract                
                infection"""

model        = OpenAI('key', model = "gpt-3.5-turbo")
nlp_prompter = Prompter(model)


result       = nlp_prompter.fit('ner.jinja',
                          domain      = 'medical',
                          text_input  = sentence, 
                          labels      = None)

eval(result['text'])

Output

[{'E': '93-year-old', 'T': 'Age'},
 {'E': 'chronic right hip pain', 'T': 'Medical Condition'},
 {'E': 'osteoporosis', 'T': 'Medical Condition'},
 {'E': 'hypertension', 'T': 'Medical Condition'},
 {'E': 'depression', 'T': 'Medical Condition'},
 {'E': 'chronic atrial fibrillation', 'T': 'Medical Condition'},
 {'E': 'severe nausea and vomiting', 'T': 'Symptom'},
 {'E': 'urinary tract infection', 'T': 'Medical Condition'},
 {'Branch': 'Internal Medicine', 'Group': 'Geriatrics'}]

• More examples: Question-Answering, Relation-Extraction, Classification, NER etc in three lines of code

• If you are still getting issues while trying out ChatGPT API, check this Google-Colab notebook

[suyin1203]

use jupyter

1. pip uninstall openai
2. close kernel (shut down kernel)
3. open notebook
4. !pip install --upgrade openai
5. its run

[hiddenchamp]

I wasted a lot of time to try to get it working. Maybe this helps someone:

AFTER updating with pip install --upgrade openai

Go here:

In pycharm go to settings > project blabla > python interpreter > click the + sign > find openai > click "Specify version: 0.27.0" VERY IMPORTANT > click install package. Restart pycharm.

It should work now.

[emilycowe]

I wasted a lot of time to try to get it working. Maybe this helps someone:

AFTER updating with pip install --upgrade openai

Go here:

In pycharm go to settings > project blabla > python interpreter > click the + sign > find openai > click "Specify version: 0.27.0" VERY IMPORTANT > click install package. Restart pycharm.

It should work now.

Do you have any idea how to get it working if you're not using Pycharm? Im having the same issue despite updating openai.

[hiddenchamp]

Do you have any idea how to get it working if you're not using Pycharm? Im having the same issue despite updating openai.

No unfortunately. I just started coding in python 1 month ago because of AI.
I guess if you are using something else, try to uninstall the current openai module and force the version number somehow.

[hanswang73]

I have the same problem, but the reason is that I copy a venv directory, so VSCode still use the python.exe in original venv directory ....