cross-posted from: https://programming.dev/post/7017319
https://d2l.ai/
The book was written originally by a group of Amazon engineers and strives to be a resource on getting started with deep learning.
Even if you have no interest in developing models, you should be aware of how they work under the hood.
For the AI enthusiasts, it makes them more interesting. For the AI doomists, it makes them less scary.
The book being online and an easy to remember URI makes this a great reference book that you can access from any device with an internet connection. You could read the whole thing with cURL if you were feeling wacky. You can also clone the repo and host it locally if you want to "own" a copy.
They are releasing a physical book in English this week, so for you collectors out there can have something for your shelves.
It's a shame that they don't include a chapter on ethics the way Deep Learning for Coders with Fastai and PyTorch: AI Applications Without a PhD does. I hope people read the etics chapter or use the ethics section of the associated free course to fill in that oversight on the Amazon Engineers' part.
Also it's a bit annoying that they address the fact that other resources exist but vaguely handwaved over the difference between their resources and the existing ones with "the others are not as up to date and don't have as much detail".
From a comment on Hacker News by cl42:
If folks are looking for free deep learning resources, here are a few more that I’ve personally vetted: https://phaseai.com/resources/free-resources-ai-ml-2024
FWIW, my recommendation with this sort of stuff is to build something alongside of the reading you are doing.
One of the things I’m exploring is building an OpenAI Gym-like project where we build neural nets to play various games. If that’s of interest to anyone, please post below!