Correcting ner.manual Annotations


I have a dataset generated as a result of several annotators utilizing the ner.manual recipe. I would now like to go back and correct any mistakes they had made while annotating the data. Is there any easy way to do this in Prodigy?

To clarify, I’d like to go back and edit the existing annotations. Not utilize ner.make-gold to correct the model’s predictions.

perhaps try
with a copy of the annotation to change all the answer to be “accept”: no

and then you can go through all of them and just “accept” the one you like, and correct the one you don’t

It'll even work without that part! :slightly_smiling_face:

Prodigy's input and output data have the same format, so you can always load examples you've exported with db-out back into a given recipe to reannotate it. ner.manual respects pre-annotated spans that are already present in the data.

Great! Thank you!

1 Like