Local llama and UE5 integration

Hi,

I’ve built a local Llama model and developed an API for it. Now, I need to integrate it into Unreal Engine 5 to create a custom offline chatbot for in-game interactions.

Has anyone done something similar or have tips for seamless integration and optimizing performance?

Thanks!

I’m using the Universal Offline LLM plugin. I’ve only been able to get Llama-2 to work so far. Currently, it’s unacceptably slow. However, it’s going to end up running on a HPC cluster and be pixel streamed. I hope the plugin gets updated to work with Llama-3.3.