Getting the proper template for function calling

[bryanhpchiang]

What is the definitive answer for how to format my prompts properly for function calling with llama3.1?

Seeing a lot of conflicting information across Twitter, the official Meta docs, Huggingface docs, and the docs from numerous inference providers.

I'm using a vLLM server and I want to create a lightweight prompt library from scratch to make sure there are no errors. Thank you!