Resources

There are resources which I find helpful for studying deep learning. They are listed below:

Deep Learning (Goodfellow, Bengio, Courville) is a classic book on deep learning. It is systematic and rigorous, especially for core ideas. It is well suited for building long-term foundations and for reading research papers. But it is not up to date with modern practice (e.g. transformers), and it offers little guidance on implementation.

Dive into Deep Learning is a modern, code-based book. It is mostly up to date and tightly integrates explanations with runnable code with different frameworks. It covers a wide range of contemporary architectures and techniques. But it is less systematic (perhaps, due to many authors) and less mathematically rigirous.

Deep Learning: Zero to Hero by Andrej Karpathy is a lecture series that builds neural networks from backpropagation to transformers from scratch. It demonstrates the implementation of major papers and core concepts in a simple manner.

Deep Learning by 3Blue1Brown focuses on visual intuition. It is excellent for forming a general understanding about deep learning. It is technically not heavy and should be treated as a conceptual support.

Learn PyTorch is a practical guide to PyTorch. It demonstrates how modern deep learning workflows are written and organized in code. It assumes prior understanding of important topics.