On-Device Function Calling in Google AI Edge Gallery- Google Developers Blog

This post announces the expansion of Google AI Edge Gallery to iOS, bringing on-device AI use cases powered by Gemma and other open-weight models to the platform. It highlights the integration of 'Mobile Actions' and 'Tiny Garden' demos, showcasing the FunctionGemma model's ability to translate natural language into function calls on-device with 270M parameters. The post also introduces in-app benchmarking for LiteRT's CPU and GPU performance, with specific metrics for the Pixel 7 Pro (1916 tokens/sec prefill, 142 tokens/sec decode). Developers are encouraged to explore demos, fine-tune FunctionGemma, and implement function calling.