Dive into Deep Learning

Zhang, Lipton, Li & Smola · 2023 · Beginner → Intermediate

D2L teaches by implementation: each concept arrives with the math, then the code that realises it. This shelf carries the foundational arc — tensors and calculus through linear models, MLPs, and convolutional networks — which is the stretch a beginner should read end to end.

Start reading39 chapters · 86,522 words · 6h 33m

Chapters 1–7 of the full book (the foundational arc).

Introduction

Introduction: overview57m

Preliminaries

Preliminaries: overview1m Data Manipulation11m Data Preprocessing4m Linear Algebra17m Calculus8m Automatic Differentiation8m Probability and Statistics29m Documentation2m

Linear Neural Networks for Regression

Linear Neural Networks for Regression: overview1m Linear Regression20m Object-Oriented Design for Implementation7m Synthetic Regression Data5m Linear Regression Implementation from Scratch8m Concise Implementation of Linear Regression8m Generalization13m Weight Decay9m

Linear Neural Networks for Classification

Softmax Regression17m The Image Classification Dataset5m The Base Classification Model3m Softmax Regression Implementation from Scratch5m Concise Implementation of Softmax Regression4m Generalization in Classification17m Environment and Distribution Shift22m

Multilayer Perceptrons

Multilayer Perceptrons: overview1m Multilayer Perceptrons13m Implementation of Multilayer Perceptrons5m Forward Propagation, Backward Propagation, and Computational Graphs7m Numerical Stability and Initialization8m Generalization in Deep Learning11m Dropout6m Predicting House Prices on Kaggle9m

Convolutional Neural Networks

Convolutional Neural Networks: overview2m From Fully Connected Layers to Convolutions12m Convolutions for Images10m Padding and Stride7m Multiple Input and Multiple Output Channels7m Pooling7m Convolutional Neural Networks (LeNet)7m

Pairs with these lessons

Python & NumPy Linear Algebra Calculus Neural Networks Convolutional Neural Networks PyTorch