What it does

This repo contains Chinese-language Jupyter notebooks covering all 20 lectures of Stanford’s introductory machine learning course (the one on OpenCourseWare / 163.com). Each chapter maps to a lecture: linear regression through SVMs, EM and factor analysis, PCA, MDPs, and policy search. Nine supplementary notebooks review linear algebra, probability, convex optimization, and Gaussian processes. The author recommends reading via Nbviewer because GitHub’s .ipynb renderer is sluggish.

The interesting bit

The notes don’t just transcribe—they restructure. Lecture titles are renamed to reflect actual teaching content rather than original names, and the format borrows from another popular Chinese ML notebook project by Jin Li (Tsinghua). It’s a study aid built by a student who clearly sat through the course, not a textbook clone.

Key highlights

• 20 chapter notebooks with direct links to specific lecture topics (SVMs with Lagrange duality, kernel methods, Gaussian discriminant analysis, etc.)
• 9 supplementary reference notebooks covering math prerequisites and advanced topics like HMMs and Gaussian processes
• Explicitly designed for Nbviewer rendering, acknowledging GitHub’s poor .ipynb performance
• Format inspired by an established Tsinghua ML notes project, suggesting iterative refinement of how Chinese students digest Western coursework
• 1,062 stars indicates genuine utility for Mandarin-speaking learners

Caveats

• Some reference notebooks (linear algebra, MATLAB intro, convex optimization overviews 1-2, Gaussian processes) lack linked .ipynb files in the README—unclear if they exist or are placeholders
• The README warns that GitHub rendering is slow, so the primary experience requires leaving GitHub entirely
• No English translation; strictly Mandarin

Verdict

Worth bookmarking if you’re a Chinese-speaking student working through classic Stanford ML and want structured, lecture-aligned companions. Skip if you need English materials or want interactive exercises beyond exposition.