fchollet/deep-learning-with-python-notebooks ? reverse-engineered prompt

Build me a clean companion notebook repo for the book Deep Learning with Python, with one Jupyter notebook per chapter and runnable code examples for the main topics from the book.

I want it to feel like the kind of project someone can open in Colab and just run through chapter by chapter, with the newer edition using Keras 3 and support for JAX, TensorFlow, or PyTorch as the backend. Please keep the notebooks focused on code and section headings, since the written explanation lives in the book itself.

Make sure the setup is straightforward for Colab and local Jupyter use, and include any notes needed for running the chapters that use Kaggle datasets or weights. If something depends on current docs or runtime setup, look it up online and keep the instructions simple for a normal reader.