The encoder–decoder approach was significantly faster than LLMs such as Microsoft’s Phi-3.5, which is a decoder-only model. “When comparing Mu to a similarly fine-tuned Phi-3.5-mini, ...

The separation of encoder and decoder components represents a promising future direction for wearable AI devices, efficiently balancing response quality, privacy protection, latency and power ...