ML extraction, maximum rows getting extracted is 100 datarows

Help AI Center
,
1

Hello everyone,

I am working on an Invoice extraction project, now I have an Invoice which has multiple pages and each page has 5 table rows containing 4 line item data each table, which means 1 page will have 20 data rows

So, I have trained sufficient amount of files and I can see that ML extractor can extract perfectly till 4th page i.e. 100 data rows (as 1 page has 20 data rows) but the 5th page it is not able to extract.

I am literally confused with this behavior so, please if anyone can help me and suggest a solution please.

**Note- I have trained with sufficient data.

Thank You!

1 Like
2

Hi there, what’s the ML package you’re using?

Run an evaluation pipeline with representative data, including documents with 5 pages. The metrics from the pipeline will give you how well the performance of your ML model is.

Based on this, you might have to simply retrain more documents.