Machine learning notes
A public learning journal by Joon Kim
I am learning machine learning by writing down the general ideas first, then implementing a small language model from scratch.
Read the foundational notes for general machine learning concepts. Then begin the active book path with Chapter 2. Chapter 1 is reading-only.
Foundational notes
General ML concepts, explanations, and my original notes.
Run the current chapter
Open a companion notebook in Colab and type the important code yourself.
Follow the book
The chapter headings follow Raschka’s book. Each notebook links to his original source notebook.
The learning path
Foundations → Working with text → Attention → GPT model → Pretraining → Fine-tuning
The bold step is the current starting point. Move forward one small section at a time.
Chapter notebooks
| Chapter | Topic | Colab |
|---|---|---|
| 2 | Working with Text Data | Open notebook |
| 3 | Coding Attention Mechanisms | Open notebook |
| 4 | Implementing a GPT Model | Open notebook |
| 5 | Pretraining on Unlabeled Data | Open notebook |
| 6 | Fine-tuning for Text Classification | Open notebook |
| 7 | Fine-tuning to Follow Instructions | Open notebook |
Short daily checklist
Use one notebook section per session:
[ ] Read the section for 20 minutes without coding
[ ] Explain the idea in my own words
[ ] Type and run the TODO code in Colab
[ ] Try one exercise or change one thing
[ ] Write: learned / unclear / next
[ ] Save the notebook to GitHubThat is a complete study session. A chapter can take several sessions.