How to use the response_format via the /completions API | OpenAI | Page 1

boreal arch Nov 6, 2023, 9:22 PM

#

Hi, I am trying to use the new response_format via the /completions API.

However I get the error:

    message: '1 validation error for Request\n' +
      'body -> response_format\n' +
      '  extra fields not permitted (type=value_error.extra)',

proven grove Nov 6, 2023, 11:06 PM

#

I am also seeing this (I presume you meant with the chat completions API). Also, I presume you are using the only model it says is supported (gpt-4-vision-preview). If that is the case, I am seeing the same issue.

I am not able to set the response_format to json_object I end up getting a 400. I am sending this parameter using the latest Python module version (1.1.0). Even if I set the parameter to text it still fails, even though that should be the default. I am using the gpt-4-vision-preview model.

proven grove Nov 7, 2023, 2:12 AM

#

Anyone figure this out?

proven grove Nov 7, 2023, 3:42 AM

#

The docs appear to have been updated and no longer list the image model as a supported model for this. So, perhaps that was the issue. I have not tried with non-image models.

#How to use the response_format via the /completions API