PredGen: Accelerated Inference of Large Language Models through Input-Time Speculation for Real-Time Speech Interaction

Re-assign community