Support structured return for LLM.

[you-n-g]

🌟 Feature Description

We have a lot of retry logic scattered everywhere to ensure structured output. However, there are already some existing tools like:

https://github.com/instructor-ai/instructor?tab=readme-ov-file#using-litellm

After we merge #564, I think it would be a great opportunity to support automatic structured output.

Motivation

1. Application scenario
2. Related works (Papers, Github repos etc.):
3. Any other relevant and important information:

Alternatives

Additional Notes