LaTeX fan translates Nielsen's neural-net classic into Chinese
A typesetter's approach to Michael Nielsen's free deep-learning textbook, rebuilt for better math and Chinese text.

What it does
This repo holds a Chinese translation of Michael Nielsen’s Neural Networks and Deep Learning, a well-known free online textbook. The author started it as personal learning notes, then formalized them in LaTeX for cleaner typography and easier cross-referencing of equations and figures. It also pulls in code samples as a git submodule.
The interesting bit
The project is explicitly “another” translation — the author borrowed from an earlier Chinese version, then re-typeset everything in LaTeX with XeLaTeX and open-source fonts (Noto CJK, Adobe Source families). The build process is documented down to the brew cask install commands.
Key highlights
- Full LaTeX source with
makebuild to PDF - Requires XeLaTeX for Chinese font support
- Uses Google Noto CJK and Adobe Source Pro fonts
- Code samples included via git submodule
- Marked “work in progress” by the author
Caveats
- “Work in progress” — completeness unclear from README
- Build dependencies are nontrivial: specific TeX Live/MacTeX versions, four font families, recursive clone
- No visible CI or pre-built releases
Verdict
Worth a look if you’re a Chinese-speaking reader who wants Nielsen’s book in print-quality PDF with living equations. Skip it if you just want to read the content online — the original English and the earlier Chinese translation are already there.