Dive into Deep Learning is a book which goes deep into deep learning. It goes all the way from simple multi-layer perceptrons (MLPs) to generative adversarial networks (GANs). The book is divided into sections and every section has multiple chapters.
I solved about 10-20% of the exercises. There are really a lot of exercises (about 3-4 per chapter). In some of the exercises I left some TODO’s and notes to myself so that I know what to revisit when I come back to what I’ve done. I have also read the Appendix: Mathematics for Deep Learning (I read it before reading the other sections).
I think this book has a good balance of theory and practice. The authors don’t go too heavy on the math and they always provide the implementation of the math concepts. There were some chapters where I have found it hard to connect the theory and practice (such as RNNs and GANs), but maybe this was just because I was not focused.
I think this is a really good book for anyone wanting to dive into deep learning.
You can find my exercise solutions in this GitHub repository.
Subscribe to my newsletter to keep abreast of the interesting things I'm doing. I will send you the newsletter only when there is something interesting. This means 0% spam, 100% interesting content.