<?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>LLM on Xuewen Yang</title>
    <link>https://xuewyang.github.io/tags/llm/</link>
    <description>Recent content in LLM on Xuewen Yang</description>
    <generator>Hugo</generator>
    <language>en</language>
    <lastBuildDate>Sun, 01 Nov 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://xuewyang.github.io/tags/llm/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Parallelism for LLM training, with numbers you can check by hand</title>
      <link>https://xuewyang.github.io/posts/parallelism/</link>
      <pubDate>Sun, 01 Nov 2026 00:00:00 +0000</pubDate>
      <guid>https://xuewyang.github.io/posts/parallelism/</guid>
      <description>Data, ZeRO, tensor, sequence, context, pipeline and expert parallelism, each worked through with a tiny example: what gets split, what moves between GPUs, and when to use which.</description>
    </item>
  </channel>
</rss>
