This repository contains repositories for running GPU accelerated LLMs on Home Assistant.
This repository contains the following add-ons
GPU accelerated llama.cpp with Vulkan and tool calling enabled.
Uses llama multiserver to serve Hugging Face models on demand.
GPU accelerated whisper.cpp using Vulkan.
Uses Wyoming Whisper.cpp for integration.
Installs OTA updates of the Sanctuary Systems hassos fork (GPU kernel patches)