<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Recursive Composition on Programmer.ie</title>
    <link>http://programmer.ie/tags/recursive-composition/</link>
    <description>Recent content in Recursive Composition on Programmer.ie</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Sat, 08 Aug 2026 13:09:00 +0100</lastBuildDate>
    <atom:link href="http://programmer.ie/tags/recursive-composition/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>The Most Important Idea in PyTorch: Recursive Composition</title>
      <link>http://programmer.ie/books/pytorch-zero-to-hero/04-chapter/</link>
      <pubDate>Sat, 08 Aug 2026 13:09:00 +0100</pubDate>
      <guid>http://programmer.ie/books/pytorch-zero-to-hero/04-chapter/</guid>
      <description>&lt;h2 id=&#34;pytorch-zero-to-hero--step-04&#34;&gt;PyTorch: Zero to Hero — Step 04&lt;/h2&gt;&#xA;&lt;p&gt;In the previous post we built a neural network using raw tensors and autograd.&lt;/p&gt;&#xA;&lt;p&gt;We had weights.&lt;/p&gt;&#xA;&lt;p&gt;We had biases.&lt;/p&gt;&#xA;&lt;p&gt;We had a forward calculation.&lt;/p&gt;&#xA;&lt;p&gt;We had a loss.&lt;/p&gt;&#xA;&lt;p&gt;We called &lt;code&gt;backward()&lt;/code&gt;.&lt;/p&gt;&#xA;&lt;p&gt;We updated the parameters.&lt;/p&gt;&#xA;&lt;p&gt;Nothing essential was missing from the mathematics of training.&lt;/p&gt;&#xA;&lt;p&gt;Now we are going to add &lt;code&gt;nn.Module&lt;/code&gt;.&lt;/p&gt;&#xA;&lt;p&gt;But I do not want to introduce it as another PyTorch class whose methods you need to memorize.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
