让草稿模型学会预测验证结果，提升大模型推理速度
Verification-Aware Training for Speculative Decoding
- 训练时模拟验证过程，用通过/拒绝模式监督草稿模型
- 平均接受长度提升11.4%，实际运行速度加快8.7%
- 适配多种模型和任务，无需改动原有结构
推测解码通过草稿模型生成候选词元，并由目标模型在单次前向传播中进行验证，以加速大语言模型的推理。验证过程是顺序进行的，一旦某位置被拒绝，其后所有位置均被丢弃。然而现有草稿模型训练方法依赖于对目标模型的逐词模仿，采用固定的位置权重策略，未能反映这一特性。本文提出验证感知训练（VAT），一种可插拔的训练框架，在每一步训练中模拟验证过程，并将由此产生的接受与拒绝模式转化为监督信号。VAT包含两个组件：(i) 验证头，一个轻量级联合训练的二分类器，用于监督草稿模型判断每个位置是否能通过顺序验证；(ii) 验证自适应加权，取代固定的权重衰减方案，将完整权重保持至每个样本首次拒绝点，并从该点重新开始衰减。VAT仅修改训练目标，可无缝叠加于现有方法之上，无需改变草稿模型结构、目标模型或推理流程。在EAGLE-3和DFlash上应用于Qwen3-4B、Qwen3-8B和LLaMA-3.1-8B，VAT在数学、代码和对话基准上均实现平均接受长度最高提升11.4%、墙钟速度最高提升8.7%，表现一致且显著。
原文摘要 · Abstract (English)
Speculative decoding accelerates large language model inference by using a draft model to generate candidate tokens, which are verified by the target model in a single forward pass. Verification proceeds sequentially and discards every position from the first rejection onward, yet existing draft training relies on token-level imitation of the target with a fixed per-position weighting that reflects neither property. We introduce Verification-Aware Training (VAT), a plug-in framework that simulates verification at every training step and turns the resulting accept and reject patterns into supervision. VAT consists of two components: (i) a verification head, a lightweight jointly trained binary classifier that supervises the draft model on whether each position survives sequential verification; (ii) verification-adaptive weighting, which replaces the fixed weighting schedule by keeping full weight up to each sample's first rejection point and re-anchoring the decay to start there. VAT modifies only the training objective, so it can be layered on top of existing methods without changing the draft architecture, the target model, or the inference procedure. Applied to EAGLE-3 and DFlash on Qwen3-4B, Qwen3-8B, and LLaMA-3.1-8B, VAT improves average acceptance length by up to 11.4% and wall-clock speedup by up to 8.7%, with consistent gains across math, code, and chat benchmarks. Code will be available at https://github.com/naver-ai/vat
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。