My requirement is to extract the data from PDF and insert into excel file. I am using AI Builder and power automate for this requirement.
Each pdf has 5 pages and the data is in 1,3 & 4 pages. First page has field information, 3 & 4 has table information (single table with 50 rows spanned into multiple pages).
The issue is after model is trained and while extracting the data, table data is extracted from 2, 3 & 4 pages instead of 3 & 4.
Page-2 has a table with different fields and I have not tagged it.
Hi @vbkr ,
If the table from page 2 is similar with the table in page 3, it can get the information from it. Tagging doesn't account for page number so when the prediction is done, the model will get information discarding the page number.
If you know that page 2 is always a page you won't need to read, you can call the model 2 times: One time with page range = 1 and a second time with page range = 3-4.
Another thing you could try is also tag the table in page 2 with a different table field but at prediction time you wouldn't consider this information.