d2l-en

Why we included this project

Dive into Deep Learning is an open-source book that treats the subject as something you learn by running code, not just by reading theory. It is written as Jupyter notebooks that combine the math, the figures, and self-contained runnable examples, so readers can execute each chapter as they go. The material spans the field, from linear algebra and the mechanics of training through convolutional and recurrent networks to attention, transformers, and modern applications in computer vision, natural language processing, and recommender systems, with the same content offered in PyTorch, TensorFlow, and JAX. That multi-framework setup makes it useful for teams comparing implementations or standardizing on one stack. The original authors and a large contributor community keep it current, and its adoption at 500 universities across 70 countries, including Stanford, MIT, Harvard, and Cambridge, shows how well it works as a bridge from tutorials to applied machine learning work.