deeplearningwithpython.io

What It Is

Deep Learning with Python Third Edition is an online book for learners of deep learning, co-authored by François Chollet, the creator of Keras, and Matthew Watson, a Google engineer and core Keras maintainer. The website provides the full third edition for free online reading, while print and ebook versions can be purchased via Manning Publications or Amazon. It is closer to an “open textbook + runnable code labs” format than a traditional live or recorded course.

Core Content and Teaching Format

The material covers machine learning, deep learning, hands-on Keras, and generative AI. The third edition explicitly adds or updates topics such as Transformers, building GPT-style large language models, and diffusion models for image generation. Its teaching style is hands-on and code-first: each chapter includes projects and code examples, written in Python and Keras, and runnable on top of TensorFlow, PyTorch, or JAX. Chapter code can be run as Jupyter Notebooks, executed in the browser through Google Colab links, and is also available on GitHub.

Instructors, Language, and Pricing

The authors’ backgrounds are a major strength of this resource: François Chollet is the creator of Keras, while Matthew Watson works on Google Gemini and contributes to the open-source deep learning ecosystem. The content is in English, so it is best suited to learners who can read technical English reasonably well. In terms of pricing, online reading is free, and purchasing the book is optional support; the website does not list specific prices. We did not find information about completion certificates, certification exams, study communities, or teaching assistant support.

Pros and Cons

The main advantages are its authoritative authorship, comprehensive structure, coverage of modern deep learning topics updated through 2025, and its effort to avoid excessive mathematical notation. It helps learners build understanding through code and intuition. The fact that it is freely available also gives it excellent value for money. Its limitations are that it is not an interactive course: there are no live Q&A sessions, assignment grading, or certificates. It also requires a reasonable level of Python ability, so complete beginners may find it challenging.

Who It’s For and Access from China

It is well suited to self-learners with a Python foundation who want to study deep learning systematically, as well as data scientists, graduate students, and researchers or engineers who want to get up to speed with Keras quickly. Access to the website itself from mainland China cannot be fully confirmed; however, supporting resources such as Google Colab, GitHub, and Amazon may face access or speed limitations, so we consider it partially restricted. If access is unstable, you can run the code locally, or consider alternatives such as Dive into Deep Learning, fast.ai, or deep learning courses on domestic Chinese platforms.