Getting Started with TensorFlow

Should've listened to the other review -- this book was definitely rushed. A solid 15-20% of the book is an overview of python and programming in general, another 20-30% covering basic math in tensorflow, and another 30% or so on a high level overview on what machine learning is. Is anyone who would buy a book on TensorFlow in need of an overview of any of these topics? The remainder of the book is dedicated to walking through examples that are readily available through the existing free resources / docs. This book was a cash grab -- don't fall for it.

This book doesn't add much to what's already available in terms of online tutorials on TensorFlow, directly from Google - and its beginning to get outdated quickly as TensorFlow is rapidly evolving, which means some of the code examples are broken now for the latest TF versions! I was disappointed that on the neural networks front, there are only three examples: (1) a single layer network, (2) an MLP, and (3) a two layer deep CNN, all trained/tested on MNIST. There isn't even one decent example on RNNs or LSTMs. Rather the book just points you to a piece of code available online, which actually seems to employ some higher-level third party library. Well, I could have found online examples for RNN/LSTM myself - why am I buying this book for? Also there is no realistic example on using TensorBoard, there's just a "Hello World!" level example which is not employed for any kind of machine learning example. I still give the book three stars, because I presume that it holds some value for a complete beginner to machine learning. It does offer quite a few non-neural network examples with TensorFlow, which while being useless for me (a PhD in ML), might be useful for a newbie.

Note that as of Tensorflow 0.7, I would have given this 5 stars. Unfortunately, the code in the book is out of date with TF 1.2 out now and examples only in python 2.7. I'm sure a python expert wouldn't care about that though. The style of the book is a "cookbook" for beginners. It gives very brief explanations for each line of code. In my case, I wanted to do regression and function approximation. This is one of two books I've seen cover this so far (both from the same publisher, but different authors). Anyways, I went to run the code and I get errors because I'm using TF 1.1. If you can fix this, maybe it will help you. I'm moving to other books to see if I can make this example work. TypeError: 'Tensor' object is not iterable

This book is terrible. The editor should be fired. There are a ton of spelling and grammar errors. It was obviously rushed to print to cash in on the popularity of subject.

I think that this book is good when you begin to learn tensorflow. If you already know about tensorflow, this book is not for you.

Terribly written, a lot of grammar and spelling mistakes. Not well structured and confusing to read at some times.

This book is a joke. It was put together very quickly and was absolutely not revised and corrected in any way. The grammar and spelling are so horrible that in some cases the sentences don't even make sense. I've seen sentences repeated twice right after an other. This is a total waste of money. A small example: (on the subject of classes in Python): "The variables and private methods are declared by convection by preceding them with two underscores." (should be "by convention"). It might seem like nitpicking. But these kinds of mistakes are very frequent and since ML is a complex subject, precision is paramount. Go check the reviews on amazon.com if you doubt me...

Libro interessante anche se un po' superficiale: la quantità di contenuto non è enorme, considerando che il libro è mirato ai primi passi mi sarei aspettato meno esempi specifici e una descrizione un po' più generale sui vari moduli / componenti. Molto è dedicato a pezzi di codice che a mio parere sono utili marginalmente, e non aiutano ad approfondire lo strumento; un capitolo intero è dedicato a un overview su python e come installare tensorflow. Il libro contiene comunque degli spunti interessanti e degli argomenti che è utile conoscere.

Code available on the net can serve you better than reading this book.

Ich kann vom Kauf dieses Buches nur abraten. Der Autor hat einfach die frei verfügbaren Tutorial-Beispiele der offiziellen TensorFlow-Webseite kopiert und daraus ein Buch gemacht. Dabei sind die Beispiele schlechter erklärt als auf der Webseite und die Abbildungen und Plots sind kopierte Screenshots in schlechter Auflösung und in Schwarz-Weiß, was besonders bei farbigen Scatter-Plots sinnlos ist. Ich war extrem enttäuscht. Das Buch ist sein Geld nicht wert.

https://www.amazon.ca/dp/1786468573/ref=cm_cr_ryp_prd_ttl_sol_0