FR-Spec: Accelerating Large-Vocabulary Language Models via Frequency-Ranked Speculative Sampling
Weilin Zhao, Tengyu Pan, Xu Han, Yudi Zhang, Ao Sun, Yuxiang Huang, Kaihuo Zhang, Weilun Zhao, YuXuan Li, Jianyong Wang, Zhiyuan Liu, Maosong Sun
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/thunlp/fr-specOfficial★ 54
Abstract
Speculative sampling has emerged as an important technique for accelerating the auto-regressive generation process of large language models (LLMs) by utilizing a draft-then-verify mechanism to produce multiple tokens per forward pass. While state-of-the-art speculative sampling methods use only a single layer and a language modeling (LM) head as the draft model to achieve impressive layer compression, their efficiency gains are substantially reduced for large-vocabulary LLMs, such as Llama-3-8B with a vocabulary of 128k tokens. To address this, we present FR-Spec, a frequency-ranked speculative sampling framework that optimizes draft candidate selection through vocabulary space compression. By constraining the draft search to a frequency-prioritized token subset, our method reduces LM Head computation overhead by 75% while ensuring the equivalence of the final output distribution. Experiments across multiple datasets demonstrate an average of 1.12 speedup over the state-of-the-art speculative sampling method EAGLE-2.