EvoSpec: Evolving Speculative Decoding via Real-Time Vocabulary and Parameter AdaptationTarget
Quick Answer
EvoSpec introduces a dynamic framework for speculative decoding in Large Language Models, achieving a 1.13x speedup over the static baseline FR-Spec on EAGLE-3 while reducing memory overhead by 27%.
Quick Take
This method adapts vocabulary and parameters in real-time, effectively addressing challenges in specialized domains like coding, law, and medicine.
Key Points
- EvoSpec employs dynamic vocabulary and parameter adaptation for real-time model evolution.
- Achieves 1.13x speedup over FR-Spec on EAGLE-3 benchmark.
- Reduces memory overhead by 27% compared to standard online adaptation.
- Utilizes context-aware mechanisms for efficient long-tail token retrieval.
- Proven effective in specialized domains like coding, law, and medicine.
Paper Resources
Article Content
From source RSS / original summaryarXiv:2605. 27390v1 Announce Type: new Abstract: Speculative decoding accelerates inference via a draft-then-verify paradigm, yet the output projection layer becomes a bottleneck as vocabulary sizes scale. While existing static pruning methods effectively reduce this overhead, they suffer from precipitous drops in acceptance rate in specialized domains or topic-switching scenarios due to their inability to capture dynamic distribution shifts.
To address this, we introduce EvoSpec, a framework that enables real-time evolution of the draft model through dynamic vocabulary and parameter adaptation. Unlike static or purely retrieval-based approaches, EvoSpec employs a context-aware mechanism that retrieves critical long-tail tokens via efficient semantic and statistical indexing. Furthermore, we propose a lightweight online alignment strategy utilizing curriculum learning to continually minimize the distributional gap between the draft and target models.
Extensive evaluations across specialized domains (coding, law, and medicine) confirm that EvoSpec overcomes the limitations of static baselines. On EAGLE-3, it achieves a 1. 13x speedup in these settings over the state-of-the-art static baseline FR-Spec, with 27\% lower memory overhead than standard online adaptation.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.