<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>DeepSeek on QingGo的碎碎念</title><link>https://qinggo.github.io/tags/deepseek/</link><description>Recent content in DeepSeek on QingGo的碎碎念</description><generator>Hugo</generator><language>zh-cn</language><lastBuildDate>Sun, 28 Jun 2026 00:00:00 +0800</lastBuildDate><atom:link href="https://qinggo.github.io/tags/deepseek/index.xml" rel="self" type="application/rss+xml"/><item><title>DSpark 深度解读：DeepSeek-V4 如何用「半自回归」把推理速度提升 85%</title><link>https://qinggo.github.io/2026/06/28/dspark-semi-autoregressive-speculative-decoding/</link><pubDate>Sun, 28 Jun 2026 00:00:00 +0800</pubDate><guid>https://qinggo.github.io/2026/06/28/dspark-semi-autoregressive-speculative-decoding/</guid><description>从数学原理到生产部署，一次讲透投机解码的核心思想与最佳实践。</description></item></channel></rss>