Closed LachsBagel closed 1 month ago
Functionary is currently part of the container
Llama 3.1 is on host via Ollama.
Let's replace both with Llama3.2 3B assuming it can handle function calling. Llama3.2 3B should reside on host via Ollama.
This should massively speed up inference.
https://github.com/danXyu is working on this
Functionary is currently part of the container
Llama 3.1 is on host via Ollama.
Let's replace both with Llama3.2 3B assuming it can handle function calling. Llama3.2 3B should reside on host via Ollama.
This should massively speed up inference.