Streaming Generation of Co-Speech Gestures via Accelerated Rolling Diffusion