ageron/handson-mlp ? reverse-engineered prompt

Build me a polished notebook based learning project for teaching machine learning in Python.

I want a set of Jupyter notebooks that walk through the basics all the way to deep learning, using Scikit Learn, PyTorch, and Hugging Face where it makes sense. Please include an index notebook that helps people browse the lessons, and make it easy to run either in Colab or on a local machine with clear setup instructions. The experience should feel like a guided course, with examples, explanations, and exercise solutions where appropriate.

Keep the notebooks organized in a sensible progression, starting from the machine learning basics and moving into neural networks, computer vision, NLP, transformers, and a few advanced topics like autoencoders, reinforcement learning, and quantization. If anything depends on current library behavior, look up the latest docs online before you wire it up.

Make sure the project is clean, reproducible, and friendly for someone who wants to learn by opening the notebooks and running them step by step.