Give ability to change retry prompt and model/provider

[aiseei]

Is your feature request related to a problem? Please describe.
For complex queries mistral-intruct/others on ollama dont work and multiple retries still dont give a parseble soln.

Describe the solution you'd like
Allow n retries on the same model else give a Callaback where in i can swicth the model/provider. ie from mistral/ollama to gpt4/Azure and prompt if needed.

Describe alternatives you've considered
None - other than custom

Additional context
Add any other context or screenshots about the feature request here.

[jxnl]

do you have a proposal for what that looks like?

[Mr-Ruben]

Would this work? (no need to change Instructor)

try:
    call_model1()
except Exception:
    try:
        call_model2()
    except Exception:
        call_model3()
        

Allow n retries on the same model

This is already possible with max_retries

I am very interested in being able to edit the 'retry prompt' as mentioned here