Tag
A developer argues that the edge AI community overlooks small, specialized models that can run locally on devices like smartphones, using a self-built offline Morse code recognition feature as an example. The project uses a sub-5 MB AI model with TensorFlow/Keras and LiteRT, and the entire pipeline from data generation to mobile integration was custom-built.
Introducing Gemma 4 12B, a multimodal model supporting audio, image, video, and text input with a novel unified architecture and a new MacOS desktop app powered by LiteRT.