<?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>Vlm on Tyler Zhu</title>
    <link>https://tylerzhu.com/blog/tags/vlm/</link>
    <description>Recent content in Vlm on Tyler Zhu</description>
    <generator>Hugo -- 0.154.3</generator>
    <language>en-us</language>
    <lastBuildDate>Sun, 17 Dec 2023 00:00:00 -0800</lastBuildDate>
    <atom:link href="https://tylerzhu.com/blog/tags/vlm/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Remarks on Spatial Localization in VLMs</title>
      <link>https://tylerzhu.com/blog/2023/12/remarks-on-spatial-localization-in-vlms/</link>
      <pubDate>Sun, 17 Dec 2023 00:00:00 -0800</pubDate>
      <guid>https://tylerzhu.com/blog/2023/12/remarks-on-spatial-localization-in-vlms/</guid>
      <description>&lt;h2 id=&#34;prelude&#34;&gt;Prelude&lt;/h2&gt;
&lt;p&gt;This all started when I oversaw this tweet from Timothee Darcet (co-first author on DINOv2)&lt;/p&gt;
&lt;p&gt;&lt;img alt=&#34;Tweet from Timothee Darcet&#34; loading=&#34;lazy&#34; src=&#34;https://tylerzhu.com/blog/images/spatial-localization/Screenshot_2023-12-16_at_12.51.07_AM.png&#34;&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href=&#34;https://x.com/TimDarcet/status/1726320282028360131?s=20&#34;&gt;https://x.com/TimDarcet/status/1726320282028360131?s=20&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;This was in response to people overreacting to how the &lt;em&gt;final&lt;/em&gt; problem in computer vision was for AI to tell the difference between a blueberry muffin and a chihuahua, which, to be fair, is a rather funny joke. It turns out that AI models can do this quite well though, and have been able to already even since CLIP came out! So what&amp;rsquo;s the big deal?&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
