@blueberry I have a question on deep-diamond. There are no public docs. Its my understanding deep-diamond is a higher level library for working with neural nets using neanderthal. Can you please confirm? Is your Deep Learning book the docs for deep diamond?
There are public docs in the docstrings (you can generate html from them using lein codox), and there are highly detailed examples in the tests (also public and free of course). The book is, of course, a very detailed guide of ins and outs (400 pages). DD is not necessarily only about Neanderthal. Neanderthal is a library for working with vectors and matrices, and related computations, while DD is a (compatible with Neanderthal) library for working with tensors, and the operations that go with that. These operations are most commonly about neural networks, but don't have to be.
thanks. How much of v2 of the dlfp book shows higher level and specific use use of deep diamond vs work with neanderthal? What about the state of v3 of the book? I'm a student/entrepreneur...haven't had income in a couple years so am taking my time on spending funds on the book.
The book builds DD from scratch, line by line, first showing how it can be done with Neanderthal, then gradually introducing tensor, and continues with the details of building a tensor library, with no characteristic details left of.
v3 will be out when the latest development of Apple silicon support is ready (not very soon, but in progress). However, I don't expect big changes, since the current book is rather stable.
My books are teaching concepts even though I insist on them going into technical details that work. They are not simple regurgiations of "here's this function, and there's that function". You can see these details from the code and the (very numerous) midje tests 😉
thanks for the detailed reply.