<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>failure detection | 木叶吟</title><link>https://yezhisheng.me/tag/failure-detection/</link><atom:link href="https://yezhisheng.me/tag/failure-detection/index.xml" rel="self" type="application/rss+xml"/><description>failure detection</description><generator>Wowchemy (https://wowchemy.com)</generator><language>en-us</language><copyright> 又拍云提供CDN服务
京ICP备16021535号-1</copyright><lastBuildDate>Tue, 31 Mar 2026 00:00:00 +0000</lastBuildDate><image><url>https://yezhisheng.me/media/icon_hu585778a5d9441f07b7d64e1beae1be58_320895_512x512_fill_lanczos_center_3.png</url><title>failure detection</title><link>https://yezhisheng.me/tag/failure-detection/</link></image><item><title>ResiHP: Taming LLM Training Failures with Dynamic Hybrid Parallelism</title><link>https://yezhisheng.me/publication/resihp/</link><pubDate>Tue, 31 Mar 2026 00:00:00 +0000</pubDate><guid>https://yezhisheng.me/publication/resihp/</guid><description/></item></channel></rss>