<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>fesenko.ai: Knowledge Base</title>
    <link>https://fesenko.ai/knowledge-base</link>
    <description>New articles on applied ML and agent engineering: where standard methods break, the known fixes, my own measurements.</description>
    <language>en</language>
    <atom:link href="https://fesenko.ai/knowledge-base/feed.xml" rel="self" type="application/rss+xml" />
    <lastBuildDate>Thu, 01 Oct 2026 00:00:00 GMT</lastBuildDate>
    <item>
      <title>Does INT8 quantization hurt small object detection? The damage mAP does not show</title>
      <link>https://fesenko.ai/knowledge-base/int8-tiny-targets</link>
      <guid isPermaLink="true">https://fesenko.ai/knowledge-base/int8-tiny-targets</guid>
      <pubDate>Thu, 01 Oct 2026 00:00:00 GMT</pubDate>
      <description>Quantizing a detector to INT8 costs small objects 1.4 to 3 times more than the overall metric, and below 12 pixels the published loss reaches a third of the accuracy; what is known, the fixes, and what nobody has measured yet.</description>
      <category>Edge deployment</category>
      <category>Research</category>
    </item>
  </channel>
</rss>
