<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0">
    <channel>
        <title>VLM - 标签 - vllbc02&#39;s blogs</title>
        <link>https://blog.vllbc.top/tags/vlm/</link>
        <description>VLM - 标签 - vllbc02&#39;s blogs</description>
        <generator>Hugo -- gohugo.io</generator><language>zh-CN</language><copyright>This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.</copyright><lastBuildDate>Tue, 28 Oct 2025 00:00:00 &#43;0000</lastBuildDate><atom:link href="https://blog.vllbc.top/tags/vlm/" rel="self" type="application/rss+xml" /><item>
    <title>patch详解</title>
    <link>https://blog.vllbc.top/patch%E8%AF%A6%E8%A7%A3/</link>
    <pubDate>Tue, 28 Oct 2025 00:00:00 &#43;0000</pubDate>
    <author>vllbc</author>
    <guid>https://blog.vllbc.top/patch%E8%AF%A6%E8%A7%A3/</guid>
    <description><![CDATA[<p><strong>Patch</strong>
就是把一张大图像切分成多个<strong>小方块</strong>（小图像块），然后把每个小方块当作一个”token”（类似文本中的单词）来处理。</p>
<h4 id="为什么需要-patch"><strong>为什么需要 Patch？</strong></h4>
<p><strong>传统 CNN</strong>：</p>
<pre><code>图像 [224×224×3] → Conv → Pool → ... → 特征</code></pre>
<p><strong>Vision Transformer (ViT)</strong>：</p>]]></description>
</item>
</channel>
</rss>
