Deep Learning with TensorFlow 2 and Keras: Regression, ConvNets, GANs, RNNs, NLP, and more with TensorFlow 2 and the Keras API, 2nd Edition

Jan 26, 2020 Review Notes 13. Some programs are giving "cublas64-100.dll" file not found error. Is it possible for authors to zip this dll file and post it on this book's Github page please? Due to fast updating libraries/tools of Python, R Programming etc 14. Source code of Packt Published books are not working. 15. If Packt publishing & authors of Packt published books can regularly test the code files and upload updated code file these books will be very useful for many years 16a. Below given book is also giving errors similar to this book's errors. 16b. Please help in fixing the source code files of the following book: "Deep Learning with R for Beginners: Design neural network models in R using TensorFlow, Keras, and MXNet by Mark Hodnett etc" have similar errors Jan 25, 2020 Review Notes: Downloaded latest update of the source code files from Github. Ran cifar10_predict.py program of page 131. It ran without errors and gave output results. However, the output gave [4 4] This output is saying both "standing cat imge" and "dog image" belong to same class of four. This result may be wrong, due to one or both of the following reasons: a. Model file "cifar10_weights.h5" used by this program is wrong? b. Accuracy of training program that generated this model file is very low? Questions are: 10a. Which is the traing program that generated the above model file? 10b. Is it the program on pages 128 and 129? 11a. Program on page 129 is saving to "model.h5" file 11b. I ran the program on page 129 and renamed the model file "model.h5" as "cifar10_weights.h5" 12a. Then I ran program on page 131 and getting following error: ValueError: You are trying to load a weight file containing 13 layers into a model with 6 layers. 12b. Authors need to fix these errors please? 12c. Fix model file names of programs on pages 129 and 131 please? Thank you. Jan 24, 2020 Review Notes: Publisher of this book has stated on this book's Github web page, that the corrections to source code files will be made in few days and posted to Github. This is an excellent and very well written book and is filled with essential information about deep learning concepts and programming techniques. This is a must to have book for persons working with Python and Deep Learning. Thanks and best regards, Jan 20, 2020 Review Notes Authors have not updated code files from Chapter 4 per my notes 1 to 6 below: They added some missing image files to chapter 4 folder. 8. Many programs from Chapter 4 are giving errors and not running. Author seems to have installed older versions of Python tools few years back when they started writing the book. Now many of the Python tools have new versions and have deprecated or removed features. Therefore many of this book's programs are not working with newer versions of Python tools. 9. Therefore, authors need to install latest versions of all the software tools on a clean new computer and test all the programs and update github web page with the source files that can be run using latest software tools versions please. Thanks. Jan 19, 2020 Review Notes As per my notes 1 to 6 below, authors have quickly posted within few days,Python Source code (missing or corrected) of Chapters 4 and 5 to the Github webpage of this book. This is great, thanks to the authors. I had only print edition of this book so far. Now I bought Amazon Kindle Fire edition of this book also. I request one more suggestion (note # 7) to the authors: 7. Please rename each source code file by prefixing with pgXXX_ corresponding to approximate page number of the code. For example, for page 115, pg115_leNet_CNN_mnist.py This kind of renumbered file names will help readers of this book, easier and quicker to find source code file and vice versa. If a source code file is discussed on multiple pages, then the file name needs to be pg115_116_117_leNet_CNN_mnist.py If authors can quickly rename the source code files and post updated source code file names to Github, then I will upgrade stars of this book from four stars to five starts please. Thanks Following notes are regarding second edition of this book: 1. Chapter 5 python source code files are missing on Github download webpage 2. Many programs in Chapter 4 are not working and giving many run-time errors 3. Python program on page 131 is not working 4. Image files are missing (in github downloaded zip file) 5. On page 131 program, getting cifar10_architecture.json can't be opened error 6. On page 131 getting error with .astype Many other python programs in the book are not working Authors are requested to quickly fix these errors and upload corrected programs and updated readme listing corrections made to the downloading zip file on the github download website please. Thanks and best regards, Authors of this book have quickly made changes within few days, per my above notes 1 to 6, and posted corrected/missing code files to the Github. Thanks to the authors for quick fixes.

This is mostly a good introductory book. Especially useful since there is a bit of a dearth on Tensorflow 2.x books out there. One thing I was interested in and was how to import keras and bring in datasets. So far I have gotten most of the way through Chapter 1. Disturbingly, on the IMDB example there are errors in the code. It can be fairly straightforward to fix them. One was in the statements to import the IMDB database. It is like the author thought of implementing the code one way, changed their mind and didn't check it out. I assume this happens later in the book too. I have coded a fair amount in tensorflow1.x and have been looking for a good book on tensorflow2.x. So here I am. I wanted one that also explored how keras was integrated into tensorflow. I know there is a script to convert files but I was worried the process would not be overly transparent, so I wanted to type in the examples myself. So far I am glad I went this route. Tensorflow2.x seems a vast improvement and is much easier to follow and this is evident in the code examples I have read and implemented so far in chapter 1. I don't mind so errors, just not too many. I think of correcting them as a form of homework. Since I am using it to self study and get up to speed on Tensorflow2.x I am happy so far. If my opinion changes as I get further into the book, I will update this review.

As a deep learning practitioner in the computer vision space, I found the academic content in Chapter 8 RNNs, Ch. 9 AutoEncoders, Ch. 11 Reinforcement Learning, Ch. 15 The Math behind Deep Learning and Ch. 12 TFX to be most useful. Of course, the other chapters are useful as well, and do tie into the ones I just mentioned. But as a professional in the field of machine learning, I got the most lift in the chapters I mentioned above. These topics will help you get a data science job or develop your current skill set if you’re already in the field: RNNs: Excellent review for those of us that work with them, and it does touch on some advanced topics like transformers, which are SOTA as of 2020. Reinforcement Learning: It goes into depth on Q-Networks and deep deterministic gradient policies, so a data scientist will understand these better when implementing from outside code tutorial -- very useful for this. Auto Encoders: Lots of business use cases can take advantage of these, so I recommend taking the time to learn this, and code them out. Math Behind Deep Learning: More for beginners entering the deep learning field, however, for experienced data scientists it wouldn't hurt to go over derivatives, differentiation, chain rule and backpropagation. TFX: A little light, it’s more of a detailed outline, I wish there was a docker-compose and coding example on how to set it up and connect TFX Pipe to TF/Keras models. External investigation is needed to develop this skill set, as getting models to prod are crucial.

I started reading the book few weeks ago. I must say it is lovely and nicely written. It is easier for to read it after being in touch with Keras, fastai(build on top pytorch). Of course, with some machine learning background things can go smoothly. My recommandation would be to dig in well the first chapter as it has the base concepts of machine learning. I do recommend it! And I love it!

Excellent book for understanding state of the art deep learning models with great code examples. Definitely worth the time to explore in full

It is a book written by someone who has a lot of knowledge about the subject, but whose ability in teaching is inversely proportional. the authors throw a lot of concepts very scarcely and badly defined, mix a lot of stuff together and don't mind spend time to explain things that are not obvious. I had several professors at college that were like this, everything is obvious to them, so they can't present things in an orderly manner. If you already have the knowledge to understand this book, it is too basic for you. Simply a bad book. Don't recommend.

Muy bueno

A very good introduction. A few typographic errors but fairly obvious to spot. Most annoying is the use of acronym without definiton such as Self Driving Car (SDC). I had to look up ontology but generally the authors describe the concepts and algoritms extremely well. Chapter 10 unsupervised learning covers PCA,K-NN and RBM and uses mathematical terms which may be unfamilar. I think a section on Bayes and PDF could be added to the variation autoencoder section (or the mathematics chapter). The best thing is lots of coding examples. For an absolute introduction I still recommend Deep Learning for Coder with fastai and Pytorch but if you have a basic knowledge this book is wider in content and great value.

The key is in the book’s title: flow. Yes, that’s my very own (100% bio/natural ;-) ) neural network eventually got to when trying to concisely describe this book. Given the non-triviality of the topics that the authors wrote about, that alone is a remarkable outcome IMHO. There’s a subtle though absolutely pragmatic approach in every chapter that guides the reader’s reasoning to a double win: grasping the inner value of the core concepts and quickly gaining real world examples (through code). I also found the vast majority of chapters to be almost ‘self consistent’: although some cornerstones are required (and thoroughly dealt with in the first few chapters) you’ll find yourself jumping back straight to, say, GANs or AutoML focused chapters for future reference or deeper dives. The ‘math focused’ chapter is an added bonus which, although not stricty necessary for the book’s mission, deserves its own credit and will give you some extra ‘Ah!’ moments.

Colour edition would have been much better for the price, as it gets bored to study this black and while, we are dealing with complex subject like deep learning, image classification n all