Deep Learning from Scratch with PyTorch
From Mathematical Foundations to Production-Ready Models
Master deep learning from mathematical foundations to production-ready models. Cover CNNs, RNNs, Transformers, GNNs, GANs, VAEs, Reinforcement Learning, and more with PyTorch implementations.
Math Foundations— Linear algebra, calculus, and probability.
Development Environment
Setting up for deep learning development
Linear Algebra for Deep Learning
The mathematical foundation of neural networks
Calculus for Deep Learning
The mathematics of optimization
Probability and Information Theory
The probabilistic foundations of learning
PyTorch Basics— Tensors, neural networks, and training.
PyTorch Fundamentals
Mastering PyTorch's core abstractions
Neural Network Building Blocks
Components of neural networks
From Perceptrons to Deep Networks
Understanding neural network architectures from shallow to deep
Data Loading and Processing
Efficient data pipelines
Backpropagation from Scratch
The algorithm that makes deep learning possible
Training Neural Networks
The art and science of training
CNNs— Convolutions and vision architectures.
Convolution Operations
The convolution operation explained
CNN Architectures
Evolution of CNN architectures
CNNs in Practice
Practical CNN techniques
Project - Image Classification System
End-to-end image classification
Sequence Models— RNNs, LSTMs, and attention.
Recurrent Neural Networks
Processing sequential data
LSTM and GRU
Gated recurrent architectures
Sequence-to-Sequence Learning
Mapping sequences to sequences
Project - Sentiment Analysis
Text classification with deep learning
Attention and Transformers Overview
Modern sequence modeling (see Transformer book for depth)
Graph Networks— GCN, GraphSAGE, and GAT.
Graphs and Graph Learning
Learning on graph-structured data
GNNs in Practice
Practical graph neural networks
Generative Models— VAEs, GANs, and diffusion.
Autoencoders
Learning representations
Generative Adversarial Networks
Adversarial training
Diffusion Models Overview
Brief introduction to diffusion (see Diffusion book for depth)
Self-Supervised— Contrastive learning and CLIP.
Self-Supervised Learning
Learning without labels
Contrastive Learning
Learning by comparison
Reinforcement— DQN, PPO, and RL in practice.
Reinforcement Learning Fundamentals
Learning from interaction
Deep Reinforcement Learning
Neural networks meet RL
RL in Practice
Practical reinforcement learning
Production— Scaling, compression, and deployment.
Advanced Optimization
Cutting-edge optimization techniques
Scaling Deep Learning
Training at scale
Model Compression
Making models smaller and faster
Deployment
From research to production
Projects— YOLO, NMT, and time series.
Project - Object Detection with YOLO
Real-time object detection
Project - Neural Machine Translation
Building a translation system
Project - Time Series Forecasting
Predicting the future
Best Practices— Professional deep learning workflows.
Deep Learning Best Practices
Wisdom from practitioners
Where the book lands in practice.
Project - Image Classification System
End-to-end image classification
Open chapter179 sections. Begin with one.
Chapter 0 — Development Environment — is where every reader starts.