Repetition for page headings

#24
by kevinqn - opened

I'm wondering if anyone else has found that the model tends to get stuck in repetitive loops when parsing metadata or other page headings. In this example, the model repeats the journal, year, and doi found in the page heading. This goes on for a while before it starts actually parsing the full text. I found the same happens in a wider variety of academic paper settings. Please let me know if you have any tips to share!

INPUT
input.png
OUTPUT
output.png

yes same here

I have observed the issue of repitition for invoices too.

Sign up or log in to comment