← All books
Book · Beginner · 35+ hours

How Neural Networks Work

A Gentle Journey from Perceptrons to CNNs, RNNs, and Beyond

Build a real intuition for neural networks from the ground up. Start with a single perceptron, work through activations, losses, backprop, and optimisers, then build CNNs for vision and RNNs for text — all with Python first and PyTorch second so the ideas stick.

20Chapters
65Sections
19hReading
7Parts
Part I·3 chapters · 11 sections

FoundationsWhat neural networks are and the tools you'll use.

What Are Neural Networks?

The big picture of neural networks

3 sections37 min read
Start chapter
  1. 01Biological Inspiration and History12m
  2. 02The Artificial Neuron15m
  3. 03Types of Neural Networks Overview10m

Python and PyTorch Essentials

Setting up your toolkit for neural network development

4 sections68 min read
Start chapter
  1. 01Python Refresher for Neural Networks15m
  2. 02Introduction to PyTorch18m
  3. 03Tensors and Operations20m
  4. 04Autograd Basics15m

Mathematics for Neural Networks

The essential math you need and nothing more

4 sections68 min read
Start chapter
  1. 01Vectors, Matrices, and Operations18m
  2. 02Derivatives and Gradients20m
  3. 03Probability Basics15m
  4. 04The Chain Rule15m
Part II·3 chapters · 9 sections

Building BlocksPerceptrons, activations, and losses.

The Perceptron

The simplest neural network

3 sections45 min read
Start chapter
  1. 01Single Neuron Model15m
  2. 02Learning Algorithm18m
  3. 03Limitations and the XOR Problem12m

Activation Functions

The non-linearity that makes networks powerful

3 sections40 min read
Start chapter
  1. 01Sigmoid, Tanh, and ReLU18m
  2. 02Choosing the Right Activation12m
  3. 03Implementing Activations in PyTorch10m

Loss Functions

How networks measure their mistakes

3 sections45 min read
Start chapter
  1. 01Regression Losses15m
  2. 02Classification Losses18m
  3. 03Custom Loss Functions in PyTorch12m
Part III·3 chapters · 10 sections

Training a NetworkForward pass, backprop, and optimisers.

Part IV·3 chapters · 10 sections

Deep NetworksMLPs, training in practice, and regularisation.

Part V·3 chapters · 9 sections

CNNs for VisionConvolutions, architectures, and an image classifier.

Part VI·3 chapters · 10 sections

RNNs for SequencesRNN, LSTM/GRU, and a text classifier.

Part VII·2 chapters · 6 sections

Going FurtherModern training techniques and debugging.

Modern Training Techniques

Tips and tricks from the pros

3 sections42 min read
Start chapter
  1. 01Batch and Layer Normalization18m
  2. 02Mixed Precision Training12m
  3. 03Gradient Clipping and Accumulation12m

Debugging and Improving Networks

When things go wrong and how to fix them

3 sections95 min read
Start chapter
  1. 01Common Training Problems28m
  2. 02Visualization Techniques35m
  3. 03Performance Optimization Tips32m
The capstone

Where the book lands in practice.

Chapter 15·3 sections

Image Classification Project

Build a complete image classifier

Open chapter
Chapter 18·3 sections

Text Classification Project

Build a sentiment analyzer

Open chapter

65 sections. Begin with one.

Chapter 1 — What Are Neural Networks? — is where every reader starts.