DFKI-NLP / sherlock

State-of-the-art Information Extraction
3 stars 1 forks source link

Introduce the concept of a "document splitter" #15

Open ChristophAlt opened 4 years ago

ChristophAlt commented 4 years ago

One frequent use case is to process a document in parts, e.g. named entity tagging or relation classification on a per sentence level. This could be for efficiency reasons or because the model was trained on a sentence level.

We need a concept that allows us to do this transparently, i.e. document -> splitter -> annotator -> combine -> document