<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Torch on Jaidev&#39;s Blog</title>
		<link>https://jaidevd.com/tags/torch/</link>
		<description>Recent content in Torch on Jaidev&#39;s Blog</description>
		<generator>Hugo</generator>
		<language>en-us</language>
		
		
		
		
			<lastBuildDate>Thu, 12 Mar 2026 13:00:00 +0530</lastBuildDate>
		
			<atom:link href="https://jaidevd.com/tags/torch/index.xml" rel="self" type="application/rss+xml" />
			<item>
				<title>Notes on Optimizing Torch Models</title>
				<link>https://jaidevd.com/posts/optimizing-torch-models/</link>
				<pubDate>Thu, 12 Mar 2026 13:00:00 +0530</pubDate>
				<guid>https://jaidevd.com/posts/optimizing-torch-models/</guid>
				<description>&lt;p&gt;ML researchers are from Mars and the ML engineers responsible for deploying&#xA;models are from Venus. The two have vastly different motivations. The ML&#xA;researcher&amp;rsquo;s job, given a dataset and some compute, is to find the&#xA;lowest possible loss on a task. In this pursuit, no engineering cost is too&#xA;high. No tech debt is too large. Worse still, if they get published, they must&#xA;include their code in their paper. Reproducible research only means that&#xA;everyone be able to reproduce benchmarks and charts from the original paper. It&#xA;usually has very little to do with production.&lt;/p&gt;</description>
			</item>
	</channel>
</rss>
