FEATURE: Ability to choose between GPT-3.5 and GPT-4 when creating a persona

[dpayne-voss]

We would like to leverage GPT-4 for certain use cases, but GPT-3.5 for the more general use cases. Right now you have to choose a specific model to be configured in the environment settings.

[fsch-ppi]

@dpayne-voss
What I did is introduce a slider, in which you can change models on the fly. I extended each ChatTreadModel by a field "model_name" and create a client with different api keys if the model is selected.

I also added a new column for images - in which the default extension of creating images is disabled. In all other modes this is disabled (since my users had trouble with images being generated that were not requested)

This does not associate models with persona, but you could do this the same way - by just adding a "model_name" to the PersonaModel