<?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>Unigram on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/unigram/</link><description>Recent content in Unigram 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>Sun, 28 Jun 2026 11:03:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/unigram/index.xml" rel="self" type="application/rss+xml"/><item><title>分词器原理与工程实践：BPE vs SentencePiece vs Unigram</title><link>https://guijiagi.com/posts/tokenizer-principles-and-practice/</link><pubDate>Sun, 28 Jun 2026 11:03:00 +0800</pubDate><guid>https://guijiagi.com/posts/tokenizer-principles-and-practice/</guid><description>深入解析BPE、WordPiece、Unigram、SentencePiece四种分词算法的原理、优缺点与2026年工程实践</description></item></channel></rss>