Answered

Document Processing Model PDF page size limitation

(0) Share

Report

Posted on by vikaskottari

Hi, I am trying to extract data from ustructured Insurance documents and each document is of around 95 pages. When I create the collection by adding 5 PDF documents and try to tag, I see only pages till 32 (sometimes only 30). I am not able to go beyond page#32 by using the down arrow or by manually typing a page number in the Go to page text box. Below is the screenshot.

Can anyone help me if this is a limitation and what is the workaround.

Screenshot 2023-07-24 130556.png

Categories:

AI Builder

I have the same question (0)

All responses (7)

Answers (1)

Sort by

CedrickB Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!

Looks like related to this thread Not able to see all the pages of document in AI Bu... - Power Platform Community (microsoft.com)

If none of the steps are working you should open a support ticket.

Thank you

Was this reply helpful? Yes No
vikaskottari 33 on at

Like
a
(0)

Report
Copy link

Link copied!
Still same issue, no luck.
The word documents that we saved as pdf were in 97-2003 document format. Now I tried upgrading all to the newest docx first and then saved as pdf; but still same issue. Able to navigate only till page 30-33.
All the pdf opens fine in browser.
Tried in Chrome and Edge browser.
But, it doesnt throow any error/warnings!

Was this reply helpful? Yes No
CedrickB Microsoft Employee on at

Like
a
(0)

Report
Copy link

Link copied!

We are sorry for the inconvience.

Can you open a support message so we can get all the necessary information to investigate please?

Was this reply helpful? Yes No
vikaskottari 33 on at

Like
a
(0)

Report
Copy link

Link copied!

We have already raised the support ticket and it has been more than a week, still no solution and the Microsoft team says they are still working on it. (TrackingID#2307250010000996)

Was this reply helpful? Yes No
Verified answer

vikaskottari 33 on at

Like
a
(0)

Report
Copy link

Link copied!

Microsoft now has confirmed that the issue has been resolved, and I can now sroll and tag till the last page of 100 page PDF.

Was this reply helpful? Yes No
vaishnavimali09 2 on at

Like
a
(0)

Report
Copy link

Link copied!

Hi. I'm Power Automate/AI builder user. I noticed that the combined length of all PDFs in a collection cannot exceed 500 pages.
That causes problems since I really want to train my model using some PDFs that are 300-450 pages long.
Is there a setting to change that would get rid of this limit? If not, does anybody have a suggestion on what they’d do as a workaround? Thanks!

Was this reply helpful? Yes No
takolota1 4,980 Moderator on at

Like
a
(0)

Report
Copy link

Link copied!

@vikaskottari

With documents that large I would use a text embedding / RAG set up to filter things down to only the relevant pages before passing the page text to a GPT prompt for data extraction:

https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Query-Large-PDFs-With-GPT-RAG/td-p/2650178

Was this reply helpful? Yes No