A voice assistant powered by Raspberry Pi 5, working completely offline!
Uses VOSK for speech-to-text, llama.cpp for LLM processing, and KittenTTS to speak!
Built in an AIY voice kit.
I used AI to help me convert from Ollama to llama.cpp
A voice assistant powered by Raspberry Pi 5, working completely offline!
Uses VOSK for speech-to-text, llama.cpp for LLM processing, and KittenTTS to speak!
Built in an AIY voice kit.
I used AI to help me convert from Ollama to llama.cpp
I added a GitHub action to compile it into a pyinstaller binary (with help from Gemini)
Log in to leave a comment
I experimented with the Kakao KANANA and LG EXAONE LLMs, as they’ve worked well in the past.
They are VERY fast even at high quantizations, but lose some world knowledge compared to Gemma3n or Phi4-mini.
Log in to leave a comment
What have i done since then?
Log in to leave a comment
Experimented with different models in llama.cpp and found Gemma 3n e2B to be the best balance of performance to world knowledge
Log in to leave a comment