<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>Speculative Decoding - 标签 - lilfry's library</title><link>https://lilfry09.github.io/tags/speculative-decoding/</link><description>Speculative Decoding - 标签 - lilfry's library</description><generator>Hugo -- gohugo.io</generator><language>zh-CN</language><managingEditor>lilfry@sjtu.edu.cn (lilfry)</managingEditor><webMaster>lilfry@sjtu.edu.cn (lilfry)</webMaster><copyright>All rights reserved.</copyright><lastBuildDate>Sat, 11 Jul 2026 23:30:00 +0800</lastBuildDate><atom:link href="https://lilfry09.github.io/tags/speculative-decoding/" rel="self" type="application/rss+xml"/><item><title>读 Qwen Bebop 与 DeepSeek DSpark：Speculative Decoding 正在从技巧变成系统工程</title><link>https://lilfry09.github.io/ai/qwen-bebop-deepseek-dspark-speculative-decoding/</link><pubDate>Sat, 11 Jul 2026 23:30:00 +0800</pubDate><author>fry</author><guid>https://lilfry09.github.io/ai/qwen-bebop-deepseek-dspark-speculative-decoding/</guid><description><![CDATA[<p>最近两篇关于 <code>speculative decoding</code> 的论文很值得放在一起看。</p>
<p>一篇是 Qwen Team 的 <strong>Bebop</strong>，全名是 <em>Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling</em>。它关心的是：在大规模 RL 后训练里，rollout 太慢，能不能用 <code>MTP</code> 把采样阶段加速起来？</p>
<p>另一篇是 DeepSeek 的 <strong>DSpark</strong>，全名是 <em>Confidence-Scheduled Speculative Decoding with Semi-Autoregressive Generation</em>。它关心的是：在线上高并发 serving 里，speculative decoding 为什么经常一上负载就不稳，怎样让它真正移动吞吐-延迟前沿？</p>
<p>这两篇看起来一个讲训练，一个讲推理；一个讲 Qwen 的 RL pipeline，一个讲 DeepSeek-V4 的线上服务。但我读完之后觉得，它们其实在回答同一个更大的问题：</p>
<p><strong>speculative decoding 的核心已经不只是“多猜几个 token”，而是怎样让猜测、验证、分布匹配和系统负载一起闭环。</strong></p>]]></description></item></channel></rss>