Deep Learning with PyTorch: Build, Train, and Tune Neural Networks Using Python Tools

Book written by deep experts, covering a lot of ground.

This book is like an extended masterclass on PyTorch from zero to competence. It develops quite beautifully in a manner that really is magnificent. I'd read it as a second book on ML after a basic course as a stepping stone from beginner to practitioner. Have read many good, bad and ugly books in the ML field - but this one deserves the title of best - it really is in a class of its own. Helped me move on and more importantly gain confidence.

The natural progression, from understanding how tensors are stored in data, all the way to a hands-on applied ML project was brilliant. Highly reccomend

- Revisit all the basics in a very ludic way - Concrete implementation for Medical Computer Vision problem (~1/3 book) - Won't make you step up in the field as intermediate/advanced user (Kaggle and papers are still the best places for that)

Testo molto chiaro scritto dai programmatori del framework pytorch.

While the content of the books published by Manning varies, the vast majority of their books are excellent, and I value their policies. I typically buy the eBook + print, starting to read and learn immediately while the paperback arrives for my collection "whenever". Better yet, Manning's books are decently priced and the publisher also provides early access to books as they are being written. As the eBook progresses and you read, you can be certain that a printed copy/final eBook arrives "when done". This is extremely important in the fast-paced topics (e.g., machine learning) that these books address, and I have recommended several books of their catalogue to students in the past. This book, as well as many others of their catalogue, are pretty much hands-on and come with complementary code examples (Manning Live Book). This provides a way to get going quickly and reproduce the examples from the book without hassle.

Book came in black and white : disappointing Book content: Really good introduction to deep learning intuition. Not particularly a fun of the last example of the book with CT scans (aka using 3d images) . An example with x-rays (2d images) scans will have been more appropriate for an introductory book, and easier to follow. Overall, very pleased

This book starts off slow, but goes into detail about PyTorch, tensors, back propagation, etc. It is a great introduction to the field and helps to understand convolutions, resnets, etc. One large basic component that it is currently lacking is a chapter on language models and attention. Hopefully the second edition will include this information down the line. Finally, the networks here are mostly sequential. The final example that takes part in the last half of the book is not incredibly useful in my opinion, but it does help to see a DL project all the way through. A few chapters about branching networks, combining 1D/2D/3D information, cross attention, and some of the current interesting complexity in the field would be welcome.

I found this book to be an excellent introduction to PyTorch. Not only is the introduction to PyTorch thorough, but its use in Deep Learning is highly documented and explained. The author doesn't scrimp on either introduction concepts or in supporting code. He spends over 475 pages to get it all spelled out carefully in text, pictures , and graphs that should satisfy the most severe critics. Python is a powerful general purpose language that has a performance bottleneck that PyTorch overcomes by accessing Nvidia GPUs to do the complex mathematical computations. Having, in effect, a Python program that can run 120 times faster than usual can make your program powerful enough to do some real research. You can design intelligent robots, self steering vehicles, house automation systems, and business research programs with this knowledge.