Nougat: Neural Optical Understanding for Academic Documents | Hacker News Reader

Nougat: Neural Optical Understanding for Academic Documents

facebookresearch.github.io

The paper (and examples) as HTML: https://facebookresearch.github.io/nougat/

Repo with code, including a CLI tool for converting a PDF to Mathpix Markdown: https://github.com/facebookresearch/nougat

We've changed to that first link from https://arxiv.org/abs/2308.13418. Thanks!

Edit: recent and related:

Nougat: Neural Optical Understanding for Academic Documents - https://news.ycombinator.com/item?id=37342215 - Aug 2023 (27 comments)

Thanks Dang! You're right that's probably the better link to use, rather than the arXiv version -- I just have a liking for PDFs, which is why I linked to the preprint :)

Trying to come to this discussion I accidentally misclicked and ended up in OPs profile, to learn that he’s a cofounder of Overleaf (hi John, big fan of the product). Overleaf, which has the purpose of turning markup language into PDFs, whereas Nougat turns PDFs into markup language. Made me wonder how many of these inverse/symmetrical/twin products there are, where one turns foos into bars and the other turning bars back into foos.

Thanks w-m, glad you find Overleaf useful :) I can't quite believe it's been ten years since we started out -- time certainly does fly by.

I did appreciate the somewhat circular loop in posting this, so thanks also for noticing! And my H/T for spotting this originally goes to my friend & colleague Alf Eaton, who reshared Lukas's tweet: https://twitter.com/lukas_blecher/status/1696101110853910716

Having this side by side with the original scan and a highlight where the text is found in the original would be an amazing product. I hated scrolling through scans of old literature in university...

New training data source for LLM just dropped

While online sources are blocking GPTBot, the real GPTBot is walking to the nearest library for the next training session.

Honestly an LLM chatbot trained on all the worlds academic writings that was able to correctly cite sources in its answers would be of immense benefit to humanity and academics.

I wish it would happen, but so much academic content is locked behind various pay gates.

How long until someone enrolls an LLM in a university course to get journal access.

tl;dr: it translates images of papers into TeX documents.

Reply on news.ycombinator.com