Slow performance - size limitation?

MeghanZiegler-1345 0 Reputation points
2024-06-21T15:53:55.06+00:00

Hi, I'm using a custom neural model to extract tables from pdfs that are about 50 pages long. One of the tables is quite large and spans 20+ pages, so the table has well over 1000 rows. The problem is that the labeling process becomes very slow after the table has 500 or so rows. After 1000, it can take a few seconds or even 20 seconds at times to insert, edit, or label a row of data in the table.

Is there a limitation to the number of rows a table can have? Or is there a Best Practice for this type of issue? I have tried editing the pdf to limit the number of pages, but this leads to poor test accuracy because each page can have uniquely different information listed in the table.

I am using the latest API version released on 2024-02-29 and have tried it on the 2023-10-31 preview as well.

Azure AI Document Intelligence
Azure AI Document Intelligence
An Azure service that turns documents into usable data. Previously known as Azure Form Recognizer.
1,500 questions
{count} votes