How fafa-dl/lhy_machine_learning works
Project type and purpose
• This is not a running application or service, but a static educational‐materials repository for 李宏毅’s Spring Machine Learning courses (2021, 2022, 2023).
• It contains year‐based directories (“2021 ML”, “2022 ML”, “2023 ML”), each subdivided by lecture topic, with Jupyter notebooks, PDFs, raw data (CSV), slide decks, and links to external platforms (Colab, Bilibili, Kaggle).
• The root README ties everything together, exposing links to videos, slides, assignments, and community channels.
File structure
• Top‐level:
• Three year folders (“2021 ML”, “2022 ML”, “2023 ML”)
• A root README.md
• Under each year:
• Folders named by lecture week or topic (“01 Introduction”, “02 Deep Learning”, … for 2021; similar naming for 2022/2023).
• Inside each topic: materials (e.g., .ipynb, .pdf), subfolders for assignments (“作业HWx”), sample code by TAs (“助教作业范例”).
• No code-build scripts, server‐side logic, or microservices—purely content.
Key insights from the README
• Points at external services:
• Bilibili (videos/slides)
• Google Colab (notebook execution)
• Kaggle, Gradescope, JudgeBoi (assignment submission/execution)
• QQ groups for community
• Maintainer’s other repos and update log—meta‐information layer.
• GitHub Repository (root)
• Year Modules: 2021 ML, 2022 ML, 2023 ML
• Content Types inside each module:
– Lecture Topics
– Assignment Folders
– Sample Code Folders
• README.md (global navigation)
• External Services: Bilibili, Google Colab, Kaggle, Gradescope, QQ Groups
• README.md → Year Modules (navigation)
• Year Module → Topic Folder → Material Files
• Material Files → External Service (e.g., notebook → run on Colab, slide → view on Bilibili)
• Assignments → Submission Platforms (Kaggle, JudgeBoi, Gradescope)
c. Architectural patterns/design principles:
• Layered content organization (Root README → Year → Topic → Material)
• Separation of concerns:
– Learning materials (notebooks/slides) vs. assignment code vs. raw data
• Decentralized content delivery: heavy reliance on external hosting (Colab, Bilibili)
d. Relevant technologies/frameworks:
• Git & GitHub for version control and hosting
• Markdown for documentation
• Jupyter Notebooks for interactive code
• PDF/slide decks for lecture materials
• External web platforms (Colab, Kaggle, Bilibili, QQ)
Checklist for the principal software engineer
• “This is a directory,”
• “This is a notebook,”
• Root README is prominent as the entry point.
• Every year‐folder and its immediate topics—don’t over‐aggregate; the value here is in the granularity of course structure.