Skip to main content
Question

Does the OpenAICompletionsConnector support the Codex model?


bruceharold
Contributor
Forum|alt.badge.img+17

I'm interested in generating SQL automatically, for use in SQLExecutor

3 replies

bruceharold
Contributor
Forum|alt.badge.img+17
  • Author
  • Contributor
  • February 10, 2023

I received an answer off line, the answer is yes, but I'm getting some laughs round the office with some responses:

2023-02-10_11-48-48


bruceharold
Contributor
Forum|alt.badge.img+17
  • Author
  • Contributor
  • February 10, 2023
bruceharold wrote:

I received an answer off line, the answer is yes, but I'm getting some laughs round the office with some responses:

2023-02-10_11-48-48

It's a Monday BTW


chrisatsafe
Contributor
Forum|alt.badge.img+2
  • Contributor
  • February 10, 2023

Hi @bruceharold​ ,

 

My understanding of the Codex model is that it is currently in limited release (beta). Personally, I haven't tried this but as long as you know the model name, you can specify any model to use in the OpenAICompletionsConnector transformer by typing it in the Model parameter.

imageThe default list of models was generated a few weeks ago but the parameter type is "choice or text" which enables users to enter their own model name in should the list fall out of sync. imageUsing the OpenAI Playground, it looks like the codex model names are code-davinci-002 and code-cushman-001 (both of which are in the default list, in addition to a few more).

 

One last thing to note, you can obtain model names using the Models endpoint with the HTTPCaller (this is how the choice or text parameter was originally populated).


Cookie policy

We use cookies to enhance and personalize your experience. If you accept you agree to our full cookie policy. Learn more about our cookies.

 
Cookie settings