Many users congratulated Sebastian Raschka on his 440-page book about building reasoning models from scratch, calling the work awesome and amazing.
No Digg Deeper questions have been answered for this story yet.
It's available from Manning (already shipping): https://mng.bz/Nwr7
and Amazon (for preorder, shipping in a few weeks): https://amzn.to/4aAKiFY
I probably would but you don't have to. It would work either way: a) You start with "Build a Large Language Model (From Scratch)" and then add inference techiques and reasoning training b) You start with "Build a Reasoning Model (From Scratch)" and then dive deeper into the foundation via "Build a Large Language Model (From Scratch)" later
@rasbt tempted to preorder. is it better to buy this first however https://a.co/d/07vPbIOl? or works either way?
Thanks, and it does! This is just an estimate, but I think it would probably around 10-15 hours of reading time, depending on whether you copy or retype the code examples. Cost-wise most of it runs on consumer hardware. For the RL chapters, I ran it on a DGX Spark. You could probably rent a A100 or H100 for $1-2 and train it in 1-5 hours. In that case, maybe <$10. There are many additional and optional experiments in the 2nd RL chapter which will take longer to run though but it's not essential.
@rasbt Nice. Just finished working through building an LLM from scratch. Want to get my copy of your new book asap.
@rasbt Congratulations Sebastian.
Bought mine.
@rasbt Congrats man!
@rasbt Ordered dead trees version - nice that it includes ebook for more instant gratification - direct from manning (hoping that gives you a better rev share than Amazon). How much time (and cost) would you estimate it would take for a motivated student to get through this?
@rasbt @philipvollet why do people doing LLM from scratch always look hot! congratulations Sebastian! 🤗
@rasbt Yes!!! Can't wait to get my hands on the printed version! Thanks for writing this super book Sebastian 🙏
@rasbt From my experience, Raschka is a significantly better educator than Andrej Karpathy. His teaching is more solid foundationally and substantially clearer.
@rasbt As a beginner I start with "Build a Large Language Model (From Scratch)". Then Reasoning Model. Correct?
@rasbt Congrats! Loved LLM from Scratch so I’ve ordered this.
It takes a certain something to write one textbook and then decide to write another one!
@rasbt finally! i pre-ordered last year! excited to receive it in India..do you know any estimates?
@MarcioK Thanks! Writing & reading are two of my favorite activities... not sure what I would do without one day 😅
@rasbt can I buy physical copy in UK? or do I have to go digital?
@rasbt Hi Sebastian, i am looking forward to this book. But I am getting error in amazon. This is for Bangalore location. Could you please help me out when is this releasing in Bangalore India ?
@rasbt Thanks for still writing books!
@rasbt Congratulations on the book!
@aaryan_kakad some links here: