<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Guidance on Yiwen Cai</title><link>https://yiwen-cai.github.io/tags/guidance/</link><description>Recent content in Guidance on Yiwen Cai</description><generator>Hugo -- gohugo.io</generator><language>zh-cn</language><managingEditor>caiyiwen.cs@foxmail.com (Yiwen Cai)</managingEditor><webMaster>caiyiwen.cs@foxmail.com (Yiwen Cai)</webMaster><copyright>© 2026 Yiwen Cai</copyright><lastBuildDate>Wed, 07 Oct 2026 11:55:07 +0800</lastBuildDate><atom:link href="https://yiwen-cai.github.io/tags/guidance/index.xml" rel="self" type="application/rss+xml"/><item><title>MIT 6.S184: Flow Matching and Diffusion Models — 第 5 章：Guidance 与无分类器引导</title><link>https://yiwen-cai.github.io/notes/diffusion/guidance/</link><pubDate>Wed, 07 Oct 2026 00:00:00 +0000</pubDate><author>caiyiwen.cs@foxmail.com (Yiwen Cai)</author><guid>https://yiwen-cai.github.io/notes/diffusion/guidance/</guid><description>从 Vanilla Guidance 的训练目标出发，用贝叶斯公式把条件向量场拆成无条件向量场加分类器梯度，推出 Classifier-Free Guidance 的线性组合形式，说明 label dropping 如何让单一网络同时学条件与无条件向量场，以及 guidance scale 只存在于采样阶段。</description></item></channel></rss>