umcu/negation-detection

Try the spacy large model to see if performance improves

Opened this issue · 0 comments

lcreteig commented

Currently using the small model. Might help a bit with tokenisation / sentence splitting.