Deep Learning with Python

I have bought 10 books on ML/DL, and of those this is the 9th book that I have read (actually I have just started reading this book, but it's been so good thus far that I wanted to write a review.) As another reviewer noted, one should read other books on ML/DI to get a deeper understanding of the topic. This book explains using programs instead of using much mathematics. The advantage that I have had is my review of the same topics from other perspectives in books such as the followingIntro to statistical learning (by Hastie et al)Intro to Machine Learning (by Alpaydin)Deep Learning (by Goodfellow, Bengio etc)Hands-on ML w SciKit, Keras and Tensorflow (by Geron)When I first tried to read this book by Chollet in early April I was not as conversant with Python, and so I took a break and decided to brush up my limited Python knowledge by going through the first 6 chapters of "Automate the Boring Stuff with Python" (by Sweigert). Now that I have more knowledge of Python this book by Chollet is so much more comprehensible. As I said I have the advantage of having learned many of these concepts earlier. I love Chollet's interpretation and explanations. I wish I could do the exercises but am having difficulty setting up the GPU machine.The problem I am dealing with with this book by Chollet is the setup of a GPU machine in the Amazon Cloud. If anyone can help me that would be greatly appreciated (I understand that this is not the forum to seek technical help on AWS, but I thought I'd give it a try)

This is probably the best into to Deep Learning one could get. Author just knows how to speak clearly, give information at the appropriate time, is well structured and still gives some very in dept info. It is limited to deep learner but that’s why its called what it is. The author dabbles in other areas so the reader is aware of other things in AI. Definitely a good starting point for someone with some programming chops but new to AI.

Read this cover to cover for my senior project and loved every minute of it, Francois Chollet was somehow able to make a textbook into a page turner, explaining challenging concepts conceptually while giving implementation examples. I also got the second addition and I would recommend using that one just so you are working through up-to-date examples with tensorflow/keras. The field of deep learning is really vast and Chollet covers an impressive amount in this book mostly at a relatively high/applied level, which I think is a good thing. There were a few of the later chapters I wish he went into more depth with, for the advanced computer vision chapter I really which he had touched on some more modern architectures like Mask- RCNN and other stuff

I'm a CS professor, and I chose this for my course in Deep Learning last term. Overall I am happy with the book, and will use it again. It rates 5 (or even 6!) stars for being an approachable introduction to Deep Learning, using the author's excellent Keras library to allow beginners to do remarkable work. My own class of undergrads was building DLNN models to do sophisticated image recognition tasks after just a few weeks.So, why the four stars? Because the book is rather "paint by the numbers". The presentation is filled with "Now you'll do this.." followed by working blocks of code for the student to enter and run. But there are no exercises, code or mathematical. Even the standard backpropagation algorithm is only qualitatively described -- nice pictures of gradient descent in 2 dimensions, but no hard equations. (After all, Keras does it all for you, right?) And as the book ventures into more advanced areas like GANs, VAEs, etc the presentation is increasingly high-level and nonmathematical, providing only a feel for the topics without deep comprehension. Given the depth of the math involved, I suppose I can't blame Chollet for a bit of handwaving. But more rigor with deeper explanations would have been nice.

If you have taken some deep learning classes on Coursera, such as deeplearning.ai or fast.ai class, this book will serve as a refresher and a good tutorial to implement ideas in Keras. While it does not provide deep theoretical concepts, it explains enough to give you an understanding of what each layer does (conv1D, conv2D, LSTM, GRU, Dense, etc.) It also teaches about different ways to assemble the networks. I especially like the chapter that talks about the functional API, where you can have multiple inputs, and multiple outputs, and layer weight sharing. Most of the other books I read only talked about Sequential models. This book is not for you, if you are looking for mathematical explanations. It's perfect for someone who is not too interested in equations, and just want to have practical understanding.

I'm using this as the primary textbook for a Deep Learning course I'm designing right now for the University of Washington professional/continuing education program. I'll also assign readings from the Goodfellow et al. text, but Chollet's book is a more practical way to get started. He is also the author of the Keras framework; it's great to get advice "straight from the horse's mouth".Overall this book is more about practical techniques and python code (in Keras) than about deep learning math/theory. This is probably what the majority of readers are looking for. It's a great synthesis of the most important techniques now (start of 2018), which is hard to get just from reading papers.I would recommend complementing this book with two others:1) as mentioned above: Deep Learning (Adaptive Computation and Machine Learning series)2) Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems