Thanks for posting the code and examples of hooking up LLaMA with WL !

Since LLaMA manuals say that its completions API adheres to the completions API of OpenAI's ChatGPT, how easy it is (or would be) to make LLaMA access configurations for use in WL's LLM* functions? (Like LLMSynthesize and LLMFunction.)

I am not totally sure now how to integrate, I mean to tie Mathematica & Wolfram Language notebook GUI to a local LLaMa. Those two local models can process images too, I can do an example in a next post. Wolfram tech usually integrates very well with other languages and databases, etc... I guess we will use in real-world projects LLMs from different sources, to attend different use cases.