"Dive into Deep Learning" (d2l-zh) is the Chinese-language edition of the widely used open-source deep learning textbook created by Aston Zhang, Zachary Lipton, Mu Li, and Alexander J. Smola. It combines clear mathematical explanations, working code, and interactive Jupyter notebooks in a single resource, letting learners read the theory and immediately run and modify the corresponding implementation.
The book progresses from foundational machine learning concepts through modern deep learning architectures, including convolutional neural networks, recurrent networks, natural language processing, attention mechanisms, and optimization techniques. Code examples are runnable and framework-based, making it easy to experiment with real models rather than just reading static text.
This resource suits learners who already have some Python and basic math background and want a rigorous, hands-on path into deep learning — from students and self-taught engineers to instructors looking for classroom material. Its "learn by doing" philosophy has made it one of the most popular open deep learning courses worldwide, adopted by hundreds of universities across many countries.
What you'll find
- End-to-end coverage from fundamentals to advanced topics like attention and transformers
- Runnable code paired with every concept, in interactive notebook format
- Free and open access, maintained by an active community