Deep Learning with Python

This is a hands on practical book for people who want to get into deep learning quickly. It requires knowledge of python but almost no knowledge of AI, explaining for instance the basic concepts of annotation, labelled instances and the difference between supervised and unsupervised learning. It starts with a series on simple practical examples which the reader can easily reproduce and explore alone. The explanations are readable and understandable away from a computer (I read much of it on holiday). It then goes into detail of the two most advanced applications of deep learning - image processing and text processing. The notation throughout is python rather than formal mathematical notation. If you like reading code but don't like reading matrix equations, this will be ideal. The one possible shortcoming is that it veers heavily to the practical side and isn't concerned with the theory. Thus it doesn't explain how backprop works or even give you the equations, merely noting that most packages automate them so you might as well not waste your time and get on with learning how to do it. This is perhaps a wise approach since Hinton's excellent coursera lectures are freely available and are both accessible and rigorous.

This book is written by someone who clearly has two major abilities: they have a love of the subject, and they communicate it clearly. The book contains real examples of Python/Keras code to do deep learning on standard data sets. Some knowledge of Python is required, but I think that any competent programmer can get this as they go along. I certainly improved my Python while working through the examples. The author makes clear their belief that a Linux system is required to do the examples in the book. This is the author's only major mistake. I have tried the examples under Windows 10/Anaconda 3 and they simply work. Perhaps the GPU based examples work better under Linux - I didn't try these. After finishing the book, the reader will be well placed to know the basics of deep learning, and to take the subject further.

This book is making something as intricate and advanced as deep learning understandable in a very clear and concise way. If you want to get started with Keras, deep learning, neural networks and all that - this is one of the best books I've ever seen. If not the best. It doesn't go full tilt into all the mathematics behind it - something I appreciate - but it sure gives you enough to get you started as well as a good way towards the more advanced subjects in this field. If you want all the formulas and algorithms behind this - there are better books but if you want to hit the ground running this is the book for you. I don't think I can recommend this book highly enough.

I'd possibly go as far as to say its the best book I've bought. I shopped around for several books and online courses on deep learning both for theory and implementing in Keras or tensorflow. None of them hit the mark. Francois Chollet wrote Keras so who better than to teach you how to use it. Even if you do not plan to learn Keras, I would still buy this for his explanation of theory alone.

This is a terrific book. It explains the topic in a clear and concise way, there's no waffle. There's real substance here and the examples are useful. The book is a nice size, I don't like weighty tomes. Reminds me of an MDX book I read years ago - I started the book in the dark and by the end I could see the light! One of the best technical books I've read, well done Francois!

Top quality book. The best AI book I have bought, with up to date explanations of what works / doesn't (mid 2017) Very well written - really explains the key concepts well. Together with the O'reilly hands on Scikit / Tensorflow book the best AI / deep learning primer to date

I've foudn this book very valuable for learning deep learning best practices and getting a feel for the field. If you learn by example and want to stay clear from math, this book is totally worth a read!

The content is clear and ideas are succinct. I am a masters AI student so to see the material written this way is great and fairly straight forward. If you are not familiar with DL at all read blogs/articles first. If you are then it will be easy to follow. However, if you want to follow along with some very simple exercises don't expect to get the same results (i.e. loss or accuracy as francis attains with the provided code). In other words he's half assed this, if he's reading this get it sorted asap mate! (very poor and frustrating for the standard at which he is working at). All of the code should be replicable within ~1% here and there as far as i'm concerned. For practical hands on experience get onto a udemy course for £15. For a quick and concise guide its okay as well - not read other books as of yet which are as accessible without heavy math and programming so from this point its well done

+ Un des livres pilliers de l'IA (ou plutôt, Deep Learning et Machine Learning) avant même la vague de mode actuelle, à lire absolument

Ótimo livro. Fiquei muito satisfeito com a compra. Linguagem simples e de boa compreensão. Único ponto negativo é que ele é todo preto e branco. Não possui figuras coloridas.

Excellent book to get a quick start on deep learning! This is not a book to learn the theoretical aspects of deep-learning, rather it is a collection of hands-on examples to work through and learn by experience and the guidance provided by the author. That said, if you have seen neural networks from the 1990s along with the back propagation algorithm, and you can visualize the concepts of gradient descent and convolution, then this material is very easy to follow The examples are setup on the Keras framework using TensorFlow as the backend engine. I used an EC2 p2.xlarge instance as suggested by the author. The setup required a bit of help beyond what's provided in Appendix B. Once setup though you will need to run from a virtual environment: "source activate tensorflow_p36". . . . . . My final thought is that after having read Chapter 7, I want to do a second pass using callbacks and tensorboard for better insight.

I like this product

Libro increíble, escrito de forma muy clara y accesible. Se lee rápido y resulta mucho más sencillo que otros textos más académicos. Aun siendo introductorio, proporciona una base tremenda para entender los conceptos fundamentales del deep learning y aprender a aplicarlos en la práctica con Keras. Ideal para quienes quieran empezar en este campo con un enfoque práctico, sin perder rigor. Muy recomendable como primer contacto antes de pasar a lecturas más avanzadas.