About This Resource

An interactive book that introduces deep learning through explanations, mathematics, and executable code. It covers neural network foundations and practical model training across several application areas.

Saved Notes

Original notes saved with this link. They may describe an earlier version.

Dive into Deep Learning — Dive into Deep Learning 1.0.3 documentation

Dive into Deep Learning

This interactive deep learning book includes code, math, and discussions.

It supports multiple frameworks including PyTorch, NumPy/MXNet, JAX, and TensorFlow.

Currently adopted at 500 universities across 70 countries.


Dive into Deep Learning (D2L) is an interactive deep learning book that has gained significant recognition in the academic and machine learning community. Here are the key highlights:

Book Characteristics

  • Implemented with multiple frameworks: PyTorch, NumPy/MXNet, JAX, and TensorFlow
  • Offers an interactive learning experience combining mathematics, figures, code, text, and discussions
  • Each section is an executable Jupyter notebook where users can modify code and tune hyperparameters

Academic Adoption The book has been adopted by over 500 universities across 70 countries, including prestigious institutions such as:

  • Stanford University
  • MIT
  • Harvard University
  • University of California, Berkeley
  • Carnegie Mellon University
  • Columbia University

Community and Learning Features

  • Interactive notebooks allow instant feedback and practical deep learning experiences
  • Enables discussion and learning with thousands of peers through community support
  • Covers diverse topics like reinforcement learning, Gaussian processes, and hyperparameter optimization

Recent Developments

  • Forthcoming publication by Cambridge University Press
  • Chinese version is a bestseller at the largest Chinese online bookstore
  • Continuous updates with new implementations and topics

The book's widespread adoption and interactive approach make it a highly regarded resource for deep learning education.

Citations: [1] https://d2l.ai