Deep Learning with Python, Third Edition

Deep learning becomes easier to approach when the ideas behind the models are connected to code you can inspect and adapt. Deep Learning with Python, Third Edition builds that connection step by step: from tensors, gradients, and model evaluation to practical work with vision, time series, language models, and generative AI.

Start with the ideas, then build with them

The opening chapters lay out what deep learning is, how neural networks learn, and why models can generalize—or overfit. From there, the book explores the mathematical building blocks and introduces TensorFlow, PyTorch, JAX, and Keras. The progression gives readers a conceptual footing before moving into larger applications.

See the methods across real tasks

Examples span classification and regression, image classification, convolutional networks, image segmentation, object detection, timeseries forecasting, and text classification. These applications show how core machine-learning choices take shape in different kinds of problems, rather than leaving the concepts abstract.

Follow deep learning into generative AI

Dedicated chapters cover language models and Transformers, text generation, and image generation. The book also addresses best practices for real-world work and closes with a look at the future of AI, placing current techniques within a broader view of the field.

Python in practice, with more than one framework

Keras is central to the book, alongside introductions to TensorFlow, PyTorch, and JAX. The examples and explanations are designed for readers with intermediate Python skills; previous experience in machine learning or linear algebra is not required.

A practical route from foundations to applications

Written by François Chollet and Matthew Watson, this third edition moves from neural-network fundamentals through a broad set of applied topics. It offers a structured way for Python programmers, students, and machine-learning practitioners to strengthen their understanding of deep learning and explore the techniques behind modern AI.