<?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>Running-Statistics on Answer</title>
    <link>https://answer.freetools.me/tags/running-statistics/</link>
    <description>Recent content in Running-Statistics on Answer</description>
    <generator>Hugo -- 0.152.2</generator>
    <language>zh-cn</language>
    <lastBuildDate>Fri, 13 Mar 2026 00:34:42 +0800</lastBuildDate>
    <atom:link href="https://answer.freetools.me/tags/running-statistics/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>归一化层的双重身份：为什么BatchNorm在训练和推理时判若两人</title>
      <link>https://answer.freetools.me/%E5%BD%92%E4%B8%80%E5%8C%96%E5%B1%82%E7%9A%84%E5%8F%8C%E9%87%8D%E8%BA%AB%E4%BB%BD%E4%B8%BA%E4%BB%80%E4%B9%88batchnorm%E5%9C%A8%E8%AE%AD%E7%BB%83%E5%92%8C%E6%8E%A8%E7%90%86%E6%97%B6%E5%88%A4%E8%8B%A5%E4%B8%A4%E4%BA%BA/</link>
      <pubDate>Fri, 13 Mar 2026 00:34:42 +0800</pubDate>
      <guid>https://answer.freetools.me/%E5%BD%92%E4%B8%80%E5%8C%96%E5%B1%82%E7%9A%84%E5%8F%8C%E9%87%8D%E8%BA%AB%E4%BB%BD%E4%B8%BA%E4%BB%80%E4%B9%88batchnorm%E5%9C%A8%E8%AE%AD%E7%BB%83%E5%92%8C%E6%8E%A8%E7%90%86%E6%97%B6%E5%88%A4%E8%8B%A5%E4%B8%A4%E4%BA%BA/</guid>
      <description>深入解析神经网络归一化层的训练推理差异：从BatchNorm的running statistics机制到LayerNorm的一致性设计，揭示为什么有些归一化层需要两个计算模式，以及工程实践中如何正确使用它们。</description>
    </item>
  </channel>
</rss>
