<?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>训练稳定性 on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/%E8%AE%AD%E7%BB%83%E7%A8%B3%E5%AE%9A%E6%80%A7/</link><description>Recent content in 训练稳定性 on 硅基 AGI · 智能体学习与测评</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Thu, 02 Jul 2026 11:02:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E8%AE%AD%E7%BB%83%E7%A8%B3%E5%AE%9A%E6%80%A7/index.xml" rel="self" type="application/rss+xml"/><item><title>神经网络归一化：LN vs BN vs RMSNorm</title><link>https://guijiagi.com/posts/neural-network-normalization/</link><pubDate>Thu, 02 Jul 2026 11:02:00 +0800</pubDate><guid>https://guijiagi.com/posts/neural-network-normalization/</guid><description>深度解析Layer Normalization、RMSNorm等归一化技术的原理、差异与演进</description></item><item><title>大模型训练稳定性：从Loss Spike到梯度爆炸的工程方案</title><link>https://guijiagi.com/posts/llm-training-stability/</link><pubDate>Tue, 30 Jun 2026 09:20:00 +0800</pubDate><guid>https://guijiagi.com/posts/llm-training-stability/</guid><description>深入分析大模型预训练过程中的稳定性问题，包括Loss Spike、梯度爆炸、训练崩溃等，并提供系统性解决方案</description></item></channel></rss>