<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom" ><generator uri="https://jekyllrb.com/" version="3.10.0">Jekyll</generator><link href="https://armangb1.github.io/pygorch/feed.xml" rel="self" type="application/atom+xml" /><link href="https://armangb1.github.io/pygorch/" rel="alternate" type="text/html" /><updated>2026-08-14T21:57:04+00:00</updated><id>https://armangb1.github.io/pygorch/feed.xml</id><title type="html">Gorch</title><subtitle>Gorch is a from-scratch automatic differentiation and neural-network library written purely in NumPy. Every backward pass is hand-written, readable, and verified against finite differences.</subtitle><author><name>Arman Gholibeikian</name></author></feed>