<?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>CLIP on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/clip/</link><description>Recent content in CLIP 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, 16 Jul 2026 11:09:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/clip/index.xml" rel="self" type="application/rss+xml"/><item><title>多模态大模型技术原理：从CLIP到原生统一架构</title><link>https://guijiagi.com/posts/b1-4a19c4f9/</link><pubDate>Thu, 16 Jul 2026 11:09:00 +0800</pubDate><guid>https://guijiagi.com/posts/b1-4a19c4f9/</guid><description>系统讲解多模态大模型的技术演进，从CLIP双塔到GPT-4V统一架构再到原生多模态训练</description></item><item><title>多模态大模型技术演进：从CLIP到原生多模态架构</title><link>https://guijiagi.com/posts/b2-19902ccc/</link><pubDate>Thu, 16 Jul 2026 10:07:00 +0800</pubDate><guid>https://guijiagi.com/posts/b2-19902ccc/</guid><description>梳理多模态大模型从早期CLIP双塔架构到GPT-4o原生多模态的技术演进路径，分析架构设计选择与工程权衡</description></item><item><title>多模态RAG：图文混合检索的架构设计</title><link>https://guijiagi.com/posts/multimodal-rag-architecture/</link><pubDate>Tue, 30 Jun 2026 09:40:00 +0800</pubDate><guid>https://guijiagi.com/posts/multimodal-rag-architecture/</guid><description>多模态RAG让AI能同时理解和检索图片与文字，本文详解四种架构设计与Claude 4/GPT-5的实践方案</description></item><item><title>多模态 RAG 实战：图文混合检索的工程实现</title><link>https://guijiagi.com/posts/multimodal-rag-image-text-hybrid-retrieval/</link><pubDate>Sun, 28 Jun 2026 10:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/multimodal-rag-image-text-hybrid-retrieval/</guid><description>深入多模态 RAG 的工程实现，涵盖图文统一编码、混合检索策略、跨模态重排序等核心技术</description></item></channel></rss>