All free AI courses
Andrej KarpathyExpertFree
Neural Networks: Zero to Hero
Build neural networks from scratch in code, one idea at a time: an autograd engine, character-level language models, and finally your own GPT and tokenizer.
9 video lessons · 18 h 38 min · taught by Andrej Karpathy
Lesson 1 of 9 · 2 h 26 min
Building micrograd: neural networks and backpropagation
Check your understanding
Course quiz
Finish the course to unlock the quiz
Mark all 9 lessons as watched and 8 questions open up here. You have 9 to go.
What you will learn
- Write backpropagation by hand and understand autograd
- Build and train character-level language models
- Diagnose training with activation and gradient statistics, and use BatchNorm
- Implement a GPT and a byte-pair encoding tokenizer from scratch
Before you start
- Prerequisites
- Solid Python and introductory calculus.
- Cost
- Free. Notebooks are on GitHub.
Videos are embedded from Andrej Karpathy’s own channel and remain theirs. The quiz is written by MTT to help you check your understanding; it is not an official Andrej Karpathy assessment.