Gemma 3n E4B It LiteRT Preview

A version of Gemma which uses Per-Layer Embeddings to speed up the inference process. This allows the model to use less than half the memory of the original model, making it suitable for edge devices and smartphones.