<?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>Transformer on s-ai-unix's Blog</title><link>https://s-ai-unix.github.io/tags/transformer/</link><description>Recent content in Transformer on s-ai-unix's Blog</description><generator>Hugo -- 0.161.1</generator><language>zh-cn</language><lastBuildDate>Fri, 30 Jan 2026 12:00:00 +0800</lastBuildDate><atom:link href="https://s-ai-unix.github.io/tags/transformer/index.xml" rel="self" type="application/rss+xml"/><item><title>AI 论文解读系列：BERT - 预训练深度双向 Transformer 的革命</title><link>https://s-ai-unix.github.io/posts/2026-01-30-bert-paper-interpretation/</link><pubDate>Fri, 30 Jan 2026 12:00:00 +0800</pubDate><guid>https://s-ai-unix.github.io/posts/2026-01-30-bert-paper-interpretation/</guid><description>深入解读 Google 发表于 NAACL 2019 的里程碑论文，剖析 BERT 如何通过双向预训练革命性地提升自然语言理解能力</description></item><item><title>AI 论文解读系列：Vision Transformer 视觉Transformer</title><link>https://s-ai-unix.github.io/posts/2026-01-30-ai-paper-interpretation-series-vision-transformer-visual-transformer/</link><pubDate>Fri, 30 Jan 2026 08:46:42 +0800</pubDate><guid>https://s-ai-unix.github.io/posts/2026-01-30-ai-paper-interpretation-series-vision-transformer-visual-transformer/</guid><description>深入解读 Google Research 的 Vision Transformer 论文，从注意力机制的原理出发，剖析图像块嵌入、位置编码、Transformer Encoder 的完整架构，揭示 Transformer 如何在计算机视觉领域挑战 CNN 的统治地位。</description></item></channel></rss>