Feature Request Thread

[zachary62]

😊 Please submit and upvote features you want

[brsolo]

Would be nice to have easier support for all OpenAI models (or mainly the latest gpt-4-turbo and gpt-3.5-turbo). Right now to use a different OpenAI model, I manually change the model parameter in init (below). I had tried to just make another elif statement, but openai.api_type is used in util.py within the openai library, so this raises an error.

One quick fix is to add the elif statement below, but it seems like bad form redefining an environ variable.

Maybe a better solution is to have two separate environ variables. One for openai.api_type and one for model.

Current code:

    elif openai.api_type == 'open_ai':
        
        response = openai.ChatCompletion.create(
            model="gpt-4-1106-preview",
            temperature=temperature,
            top_p=top_p,
            messages=messages
        )

Potential quick fix:

  elif openai.api_type == 'gpt-3.5-turbo-0125':

       os.environ['OPENAI_API_TYPE'] = 'open_ai'
       
       response = openai.ChatCompletion.create(
           model="gpt-4-1106-preview",
           temperature=temperature,
           top_p=top_p,
           messages=messages
       )