nndl-practice

Why we included this project

A PyTorch companion to Qiu Xipeng's deep learning practice book, this repo carries runnable notebooks for each chapter, covering linear models, feed-forward and convolutional networks, RNNs, attention, graph networks, and large language models. Every chapter is a self-contained notebook with a short README on implementation choices and pytest sanity checks, so the code actually runs instead of leaving you to guess. That makes it a solid study guide to keep beside the theory book, and a quick reference when you want one clean implementation of a familiar architecture. The legacy/ directory keeps the numpy-era exercises from the first edition, so people following the older material still have a starting point. Instructors can also adapt the notebook-plus-test structure into course assignments.