📘 Dive into Deep Learning – My Learning Journey

This repository documents my personal journey through the Dive into Deep Learning book. I'm working through the chapters using PyTorch, implementing core deep learning models from scratch, experimenting with the concepts, and probably extending them into small side projects. To get a deeper understanding I am not using the d2l package but implement my own Python files and packages while working through the book.

My goal is to build a strong, intuitive, and hands-on understanding of modern deep learning – from basics to advanced models – as part of my master's degree in computer science with a focus on machine learning and data science.

---

📚 What's Inside

✅ Chapters and Exercises

Each chapter folder contains:

• one notebook per subchapter where I code along and work on the exercises
• a readme with the contents of the chapter and subchapters

Folder	Chapter	Status
chapter_01_introduction	1. Introduction	✅ Done
chapter_02_preliminaries	2. Preliminaries	✅ Done
chapter_03_linear_neural_networks_for_regression	3. Linear Neural Networks for Regression	✅ Done
chapter_04_linear_neural_networks_for_classification	4. Linear Neural Networks for Classification	✅ Done
chapter_05_multilayer_perceptrons	5. Multilayer Perceptrons	✅ Done
chapter_06_builders_guide	6. Builders’ Guide	✅ Done
chapter_07_convolutional_neural_networks	7. Convolutional Neural Networks	✅ Done
chapter_08_modern_convolutional_neural_networks	8. Modern Convolutional Neural Networks	✅ Done
chapter_09_recurrent_neural_networks	9. Recurrent Neural Networks	✅ Done
chapter_10_modern_recurrent_neural_networks	10. Modern Recurrent Neural Networks	✅ Done
chapter_11_attention_mechanisms_and_transformers	11. Attention Mechanisms and Transformers	🛠 In Progress
chapter_12_optimization_algorithms	12. Optimization Algorithms	🔜 Planned
...	...	...

🛠️ Projects

Experiments or fun side projects that go beyond the book; currently none might add later, e.g.:

• cnn_fashion_classifier/: A CNN trained on FashionMNIST
• bandname_generator/: An RNN that invents indie band names 🎸
• using RL to train a model to play 3D Pinball Space Cadet

---

💡 Why I'm Doing This

I want to:

• Deeply understand the math and mechanics of deep learning.
• Strengthen my PyTorch, Python and project structuring skills.
• Build a solid foundation before diving further into NLP and LLMs (e.g. via the Hugging Face Course on LLMs).

---

🧰 Tools & Stack

• Python 3.x / 3.13.3
• PyTorch
• Numpy
• Pandas
• Matplotlib
• torchvision
• Jupyter Notebooks
• VS Code

---

📎 Resources

• 📘 Dive into Deep Learning
• D2L GitHub Repo
• Hugging Face Course on LLMs

---

📜 License

MIT