A Chinese ML textbook with 19K stars and no chapters in sight

What it does

The repository hosts metadata and build plumbing for 邱锡鹏’s Neural Networks and Deep Learning (蒲公英书 / “Dandelion Book”), a Chinese machine-learning textbook series. It tracks a v2 theory edition, a general-education spinoff, and an archived v1 PDF release. The actual chapter text, exercises, and solutions are not in this public repo; v2 and the general-edition are slated for 2026 publication.

The interesting bit

The repo’s real product is a _meta.yml convention: each book directory drops a metadata file, and a separate site-builder repo scrapes them all into a unified catalog. It’s a book-publishing pipeline disguised as a GitHub repo.

Key highlights

• Nearly 19K stars for what is essentially a bibliography card and a landing page
• v1 full materials (PDFs, slides) live on a dedicated branch, not main
• Companion repos exist for PyTorch/Paddle practice problems and a separate LLM beginner track
• Bilingual branding: “蒲公英书” (Dandelion Book) for the v2 theory edition
• Star-history chart included, presumably so visitors know other visitors know

Caveats

• v2 and general-edition content is not yet available; 2026 is the stated ship date
• README descriptions of v2 contents differ slightly between the two provided snapshots (one says “metadata and landing page,” the other says “chapters, exercises, errata”), so the actual current state is unclear

Verdict

Worth a bookmark if you read Chinese and want a structured ML curriculum once it ships. If you’re hunting for runnable code or draft chapters today, you’ll leave hungry.