一作ACL 2026Findings· Findings of ACL
SPIDE : Serial and Parallel Intertwined Speculative Decoding
Wenru Xu · Peixuan Xu · Ziqi Yang · Ming Hu · Zihui Wang · Jianzhong Qi · Rongshan Yu · Xiaoliang Fan · Cheng Wang
摘要
Speculative Decoding (SD) reduces inference latency for Large Language Models (LLMs) by leveraging an efficient draft model to generate candidate tokens, which are subsequently verified by the target model. To enhance acceleration while reducing the LLM usage costs, we propose Serial and Parallel...