Add notebook to demonstrate small LLM usage

[kellyaa]

Why are these changes needed?

Most AG2 examples do not work well with small LLMs (as stated in the docs). However, AG2 can be used with small models given the right techniques. This notebook demonstrates a tactic that can be used to get high performance out of small models in a dynamic RAG workflow.

Related issue number

n/a

Checks

• I've included any doc changes needed for https://docs.ag2.ai/. See https://docs.ag2.ai/docs/contributor-guide/documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[kellyaa]

Ready for review!

[qingyun-wu]

Thank you Kelly! Could you sign the CLA via this check:

[qingyun-wu]

LGTM! Thank you!

[davorrunje]

Please fix the spelling errors:

Idenfiy documents to ingest

The Worflow

[CLAassistant]

All committers have signed the CLA.

[kellyaa]

@davorrunje Ah, thank you

[marklysze]

Thanks @kellyaa, love having more on local and smaller models!

Have you had a chance to try it using the AG2 Ollama client instead of the AG2 OpenAI client?

e.g. "api_type"="ollama" and using the client_host for the path?

[
    {
        "model": "llama3.1",
        "api_type": "ollama",
        "client_host": "http://192.168.0.1:11434"
    }
]

[kellyaa]

@marklysze
Works like a charm! Updated the code to use this. Thank you!