<?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>Context Parallelism on Yunsheng Ni</title><link>https://niyunsheng.github.io/tags/context-parallelism/</link><description>Recent content in Context Parallelism on Yunsheng Ni</description><generator>Hugo -- 0.154.5</generator><language>en-us</language><copyright>Content is licensed under CC BY-NC-SA 4.0.</copyright><lastBuildDate>Sat, 21 Mar 2026 15:41:55 +0800</lastBuildDate><atom:link href="https://niyunsheng.github.io/tags/context-parallelism/index.xml" rel="self" type="application/rss+xml"/><item><title>Loss Reduction in Distributed Training</title><link>https://niyunsheng.github.io/loss-reduction-dp-cp/</link><pubDate>Sat, 21 Mar 2026 15:41:55 +0800</pubDate><guid>https://niyunsheng.github.io/loss-reduction-dp-cp/</guid><description>An analysis of how Data Parallelism (DP) and Context Parallelism (CP) affect loss reduction, and how to maintain mathematical equivalence between distributed and single-device training for LLM and Video DiT models.</description></item></channel></rss>