Lior Sinai • 7/27/2024

MicroGrad.jl: Part 1 ChainRules

This article is the first part of a series on creating a minimal automatic differentiation (AD) package in Julia called MicroGrad.jl. It explains the fundamentals of AD, contrasts object-based Python approaches with Julia's functional and dispatch-based system (using Zygote.jl), and details the implementation of explicit chain rules. The series aims to provide comprehensive tutorials on AD in Julia, requiring a background in Julia and Calculus.

0 comments

#Machine Learning #julia #Backpropagation