<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>ONNX | Uzziel Perez</title><link>https://www.uzzielperez.com/tags/onnx/</link><atom:link href="https://www.uzzielperez.com/tags/onnx/index.xml" rel="self" type="application/rss+xml"/><description>ONNX</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Mon, 25 May 2026 09:00:00 +0700</lastBuildDate><image><url>https://www.uzzielperez.com/media/icon_hu7729264130191091259.png</url><title>ONNX</title><link>https://www.uzzielperez.com/tags/onnx/</link></image><item><title>Optimizing GNNs for the Wild: PyTorch-to-ONNX Acceleration of GarNet on CPUs and GPUs</title><link>https://www.uzzielperez.com/event/chep2026-onnx-optimization/</link><pubDate>Mon, 25 May 2026 09:00:00 +0700</pubDate><guid>https://www.uzzielperez.com/event/chep2026-onnx-optimization/</guid><description/></item><item><title>UzzieNet / ecal-reco-gnn</title><link>https://www.uzzielperez.com/project/uzzienet/</link><pubDate>Mon, 25 May 2026 00:00:00 +0000</pubDate><guid>https://www.uzzielperez.com/project/uzzienet/</guid><description>&lt;p>Core contributor to &lt;strong>ecal-reco-gnn&lt;/strong>, the LHCb calorimeter reconstruction package that hosts &lt;strong>UzzieNet&lt;/strong>: an attention-enhanced, node-centric graph neural network extending GarNet. Custom graph construction, message-passing and attention layers, model compression, and ONNX deployment target the PicoCal (Upgrade II) calorimeter and the GPU-based HLT1/Allen trigger.&lt;/p>
&lt;p>Presented at &lt;a href="https://www.uzzielperez.com/event/acat2025-lightweight-gnns/">ACAT 2025&lt;/a>, &lt;a href="https://www.uzzielperez.com/event/fastml2025-lightweight-gnns/">Fast ML 2025&lt;/a>, and &lt;a href="https://www.uzzielperez.com/event/chep2026-lightweight-gnns/">CHEP 2026&lt;/a>.&lt;/p></description></item></channel></rss>