Keep a Raspberry Pi AI chatbot responsive by preloading the LLM and offloading with Docker, reducing first reply lag for ...