I love how that site offers a phone-formatted PDF as an option. How thoughtful!

i got duo to increase my reading.

Is it working out well for tech pdf's and books? Once I saw the form factor I realized it would be better than an iPad (less heavy, better battery) and that is the use case I am looking at. Might wait till v2.

[delayed]

[delayed]

You may enjoy the reflow feature on Koreader. It fits a pdfs content into a screenwidth single column layout. That text is also zoomable.

Does it work on scanned (non-“digital”) PDFs as well? That has been my biggest pain point so far. Haven’t tried Koreader yet.

I have found that it works surprisingly well on PDFs ripped from the Internet Archive. These PDFs tend to be OCR'd scanned copies of the book. I doubt that it would work if the text is not selectable at all.

It's a nice treatment of the subject, but I wish he would post a conventionally-formatted .PDF that puts every on a full sheet of standard printer paper.

I immediately wanted the same thing, so I made it. The book is licensed Creative Commons BY-NC-SA 4.0 International License, so here you go.

https://cdn.filepost.dev/file/filepost/uploads/15/1578f21219...

Thanks!

Thanks.

But fyi for others this is one version behind and missing these updates:

  V1.3 (May 12, 2026)
  Many typos and phrasing fixes with the help of LLMs.
  Section 3.8. Added a section about large-scale training

Oh, that's unfortunate! I made it from the A5 version on the site which I downloaded right before making my post. Looks like the author hasn't updated that one.

I just emailed him. Fingers crossed.

I'd note that the Fleuret books gives the foundation for basic deep learning whereas the OP seems to give the necessary foundations for Transformers and related machinery. These later addition what is now "AI" are in total at least as complicated as the original deep learning concepts. Not that just learning basic neural networks but the two book don't appear to cover the same ground.

Only had a quick skim, but looks like the content of this book has very little to do with transformers. Might be more relevant for diffusion models (image/video generation) though.

Gee, I think you're right. But it's definitely a leap beyond what was deep learning a while back.

The funny part is that these two worlds (diffusion and transformers) are starting to merge now - for example, in DiT , where transformer architecture is used for noise prediction. So the foundation from this book will still come in handy

My full stack JS jockey buddy studied this in the nascent days of LLMs to prep for an interview and is now generationally wealthy. Sliding doors moment for sure...