Persistent Internal Server error with GPT4-v

Question

Persistent Internal Server error with GPT4-v

Nirmal Thacker 25

I'm receiving a persistent 500 Internal Server error while trying to run a simple gpt4-v example with the computer vision enhancements.

See attached screenshot for sample code that fails for me. The editor here doesnt appear to show python code correctly formatted

Screenshot 2024-02-16 at 5.36.30 PM.png Im using Python 3.11.6 with openai==1.12.0

I get the following error and its been persistent for the last 8 hours.

openai.InternalServerError: Error code: 500 - {'error': {'code': 'InternalServerError', 'message': 'The service is temporarily unable to process your request. Please try again later.', 'param': None, 'type': None}} If I change the above to simply skip extra_body it works fine

Saurabh Sharma 23,866 Microsoft Employee Moderator

Hi @Nirmal Thacker
Welcome to Microsoft Q&A! Thanks for posting the question.

I have tried GPT-Vision model and it worked as expected. Please can you try with the below code -

import os
from openai import AzureOpenAI

api_base = 'https://{}.openai.azure.com/' # your endpoint should look like the following https://YOUR_RESOURCE_NAME.openai.azure.com/
api_key=""
deployment_name = 'gpt4vision'
api_version = '2023-12-01-preview' # this might change in the future

client = AzureOpenAI(
    api_key=api_key,  
    api_version=api_version,
    base_url=f"{api_base}openai/deployments/{deployment_name}/extensions",
)


response = client.chat.completions.create(
    model=deployment_name,
    messages=[
        { "role": "system", "content": "You are a helpful assistant." },
        { "role": "user", "content": [  
            { 
                "type": "text", 
                "text": "Describe this picture:" 
            },
            { 
                "type": "image_url",
                "image_url": {
                    "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"

                }
            }
        ] } 
    ],
    max_tokens=2000 
)
print(response)

In which region your resource is? Thanks Saurabh

Nirmal Thacker 25 Reputation points

2024-02-17T01:36:56.4166667+00:00

Screenshot 2024-02-16 at 5.36.30 PM.pnghi Saurabh, My resource is in US West For some reason the editor here doesnt show my test python code but I am continuing to hit errors when i try to use gpt4vision with the AzureComputerVision enhancements. Attached a test python code here that fails with 500 internal server error for me. When I test without the AzureComputerVision enhancements extra_body attribute it works fine for me too, but I need this feature as well.
Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-20T20:32:09.7466667+00:00

Nirmal Thacker Thanks for sharing. I will look into this and get back to you. Thanks Saurabh
Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-20T22:47:49.1333333+00:00

@Nirmal Thacker I have checked with Computer Vision enhancements as well but not getting any errors and code is working as expected.

As mentioned in the documentation, can you please check if your Computer Vision resource is in the paid (S1) tier and in the same Azure region as your GPT-4 Turbo with Vision resource. Please let me know if you still see any issues. Thanks Saurabh
Nirmal Thacker 25 Reputation points

2024-02-21T04:26:13.8466667+00:00

Hi Saurabh, my CV resource is in S1 tier Both CV and OAI endpoint are in US West The playground works, but the code is still hitting a 500 Internal Server Error Please advice how we can resolve this, really need this capability and 500 Internal server error is kinda unhelpful

Answer accepted by question author

2 additional answers

Your answer

Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-17T01:23:43.1666667+00:00

Hi @Nirmal Thacker
Welcome to Microsoft Q&A! Thanks for posting the question.

I have tried GPT-Vision model and it worked as expected. Please can you try with the below code -

import os from openai import AzureOpenAI api_base = 'https://{}.openai.azure.com/' # your endpoint should look like the following https://YOUR_RESOURCE_NAME.openai.azure.com/ api_key="" deployment_name = 'gpt4vision' api_version = '2023-12-01-preview' # this might change in the future client = AzureOpenAI( api_key=api_key, api_version=api_version, base_url=f"{api_base}openai/deployments/{deployment_name}/extensions", ) response = client.chat.completions.create( model=deployment_name, messages=[ { "role": "system", "content": "You are a helpful assistant." }, { "role": "user", "content": [ { "type": "text", "text": "Describe this picture:" }, { "type": "image_url", "image_url": { "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/2560px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg" } } ] } ], max_tokens=2000 ) print(response)

In which region your resource is? Thanks Saurabh
Nirmal Thacker 25 Reputation points

2024-02-17T01:36:56.4166667+00:00

Screenshot 2024-02-16 at 5.36.30 PM.pnghi Saurabh, My resource is in US West For some reason the editor here doesnt show my test python code but I am continuing to hit errors when i try to use gpt4vision with the AzureComputerVision enhancements. Attached a test python code here that fails with 500 internal server error for me. When I test without the AzureComputerVision enhancements extra_body attribute it works fine for me too, but I need this feature as well.
Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-20T20:32:09.7466667+00:00

Nirmal Thacker Thanks for sharing. I will look into this and get back to you. Thanks Saurabh
Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-20T22:47:49.1333333+00:00

@Nirmal Thacker I have checked with Computer Vision enhancements as well but not getting any errors and code is working as expected.

As mentioned in the documentation, can you please check if your Computer Vision resource is in the paid (S1) tier and in the same Azure region as your GPT-4 Turbo with Vision resource. Please let me know if you still see any issues. Thanks Saurabh
Nirmal Thacker 25 Reputation points

2024-02-21T04:26:13.8466667+00:00

Hi Saurabh, my CV resource is in S1 tier Both CV and OAI endpoint are in US West The playground works, but the code is still hitting a 500 Internal Server Error Please advice how we can resolve this, really need this capability and 500 Internal server error is kinda unhelpful

Answer 1

Saurabh Sharma 23,866 Microsoft Employee Moderator

Hi @Nirmal Thacker , As per internal email conversation, this issue seems to be resolved by itself. Please let us know in case you encounter it again.

Thanks

Saurabh

Please 'Accept as answer' and Upvote if it helped so that it can help others in the community looking for help on similar topics.

0 comments

Answer 2

Wilson Wang 0

Similar issue - ever since about 3-4pm PST, our gpt-4-0125-preview deployment is throwing a 500 Internal Server error on every request. I tried hitting it via API and also the Azure AI Studio playground, both failed.

Nirmal Thacker 25 Reputation points

2024-02-17T02:17:36.96+00:00

The playground seems to work for me , but unfortunately the code in the playground doesnt use the openai library in python and runs with through the requests module so im not entirely sure how that is processed. I do need to use the openai library though
Mayank Makwana 5 Reputation points

2024-02-17T03:07:27.4333333+00:00

Also experiencing this same issue.
Saurabh Sharma 23,866 Reputation points Microsoft Employee Moderator

2024-02-22T00:32:29.8266667+00:00

Nirmal Thacker I am looking into this and get back to you. Thanks Saurabh

Answer 3

Charlie Wei 3,335

Hello @Nirmal Thacker ,

I have reviewed the code screenshot you provided, and I believe the API version portion may need to be modified as follows. For more information about API versions, please refer to the official Microsoft GitHub.

api_version = "2023-12-01-preview"

Best regards,
Charlie

If you find my response helpful, please consider accepting this answer and voting 'yes' to support the community. Thank you!

Nirmal Thacker 25 Reputation points

2024-02-17T01:58:25.17+00:00

Hi @Charlie Wei just tried with api_version = "2023-12-01-preview" and still hit the error: openai.InternalServerError: Error code: 500 - {'error': {'code': 'InternalServerError', 'message': 'The service is temporarily unable to process your request. Please try again later.', 'param': None, 'type': None}}

Share via

Persistent Internal Server error with GPT4-v

2 additional answers

Your answer