Invoice node classification / meta-data extraction / single prediction with trained model #72
Description
Hi i'm Damien !
I'm doing an internship for my last bachelor's year in software developper studies, the project of my internship is doing meta-extraction from documents using machine learning.
I'm actually trying to implement something similar to this : https://github.com/udipbohara/info_extraction_receipts, but for my case, i would like to treat invoices.
How can i train a model with n invoices in entry, save it and then reload it to make a single prediction. I would like to use this trained model to put an invoice in entry and get possible labelled zones like date, amount, ...
Is it possible to use a trained model for a single prediction ? If yes, can you explain me ?
Thank you !