datawhalechina/llm-cookbook
Andrew Ng’s LLM courses, repackaged for Chinese developers
Localizes Andrew Ng’s LLM curriculum into runnable Chinese notebooks with prompts re-engineered for Mandarin.
Velocity · 7d
+3.0
★ / day
Trend
↘cooling
star history
What it does
The project takes 11 of Andrew Ng and OpenAI’s joint DeepLearning.AI courses—covering prompt engineering, RAG, LangChain agents, fine-tuning, and evaluation—and translates them into Chinese. It doesn’t stop at subtitles: the team re-created the example code as runnable Jupyter notebooks and re-engineered the prompts so they work in Chinese contexts, not just as literal translations. The result is a gated curriculum of required fundamentals and elective deep dives, served up as online docs, PDFs, and notebooks.
The interesting bit
The maintainers actually A/B tested Chinese versus English prompts to find “roughly equivalent” Mandarin versions, treating localization as a prompt-engineering problem rather than a translation problem. That attention to linguistic nuance is rare in course ports.
Key highlights
- Full coverage of the Ng/OpenAI stack: Prompt Engineering, ChatGPT API systems, LangChain, RAG, Gradio, W&B tracing, and fine-tuning via Lamini.
- Curriculum is split into required baseline courses and electives covering advanced retrieval, agents, model evaluation, and more.
- All notebooks are bilingual and runnable; the team also provides a video with Chinese subtitles for the flagship Prompt Engineering course.
- Hosted for domestic access: online HTML docs, a PDF release, and Bilibili videos—bypassing the English-only originals’ regional restrictions.
- Community-maintained by Datawhale; they accept PRs for missing Ng courses and slot them into the leveling system.
Caveats
- The code assumes you have an OpenAI API key and are comfortable with Jupyter; other LLM APIs require manual tweaks.
- The README notes the prompt equivalence is “roughly” achieved, so edge cases in Chinese understanding may still differ from the English originals.
Verdict
Worth bookmarking if you are a Mandarin-speaking developer with basic Python who wants a structured, project-based on-ramp to LLMs. Skip it if you already finished the English originals and don’t need Chinese prompt adaptations.
Frequently asked
- What is datawhalechina/llm-cookbook?
- Localizes Andrew Ng’s LLM curriculum into runnable Chinese notebooks with prompts re-engineered for Mandarin.
- Is llm-cookbook open source?
- Yes — datawhalechina/llm-cookbook is an open-source project tracked on heatdrop.
- What language is llm-cookbook written in?
- datawhalechina/llm-cookbook is primarily written in Jupyter Notebook.
- How popular is llm-cookbook?
- datawhalechina/llm-cookbook has 24.4k stars on GitHub and is currently cooling off.
- Where can I find llm-cookbook?
- datawhalechina/llm-cookbook is on GitHub at https://github.com/datawhalechina/llm-cookbook.