PARD: Accelerating LLM Inference with Low-Cost PARallel Draft Model Adaptation

Re-assign community