Issues related to model quality

Case 1: model failed to extract data

Check the document structure and compare it with that of a successfully processed one, verify the records in the training set. If it's a new document type, it's expected behavior. Add 10 similar tagged documents or so to the training set and retrain the model.

Case 2: model extracted data but failed to group fields

By default, a model can group fields only if they are in a table. Otherwise, add custom post-processing. For details, refer to the following topics:

• Post-Processing examples

• Start from Archetype

Read also:

• Issues related to training BP failure

• Issues related to Eval job failure

• Issues related to bean creation error

• Issues related to errors during field extraction