LiteRT.js runs machine learning models locally with CPU, GPU and emerging NPU acceleration, potentially reducing server infrastructure, inference charges and data movement.
With residents increasingly using AI assistants for public services, engineering head David Holmes explains how the ...