<?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>🧠 TransFuser on Elon&#39;s AD Insight</title>
    <link>https://auto-driving-blog.vercel.app/tags/-transfuser/</link>
    <description>Recent content in 🧠 TransFuser on Elon&#39;s AD Insight</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Sun, 19 Jul 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://auto-driving-blog.vercel.app/tags/-transfuser/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>论文精读｜MVAdapt：零样本多车型适应的端到端自动驾驶</title>
      <link>https://auto-driving-blog.vercel.app/posts/paper-reading/%E8%AE%BA%E6%96%87%E7%B2%BE%E8%AF%BB-2604-11854/</link>
      <pubDate>Sun, 19 Jul 2026 00:00:00 +0000</pubDate>
      <guid>https://auto-driving-blog.vercel.app/posts/paper-reading/%E8%AE%BA%E6%96%87%E7%B2%BE%E8%AF%BB-2604-11854/</guid>
      <description>MVAdapt 揭示了端到端自动驾驶的&amp;rsquo;车辆域差距&amp;rsquo;：模型隐式绑定了训练车的动力学特性，换车即翻车。它用轻量物理编码器编码轴距、质量、驱动方式等物理属性，通过交叉注意力注入冻结的场景编码器，在不损失视觉泛化能力的前提下实现跨车型适应。在 58 种车型上的零样本迁移表现大幅领先 naive transfer 和 URMA 等基线。</description>
    </item>
  </channel>
</rss>
