I'd like to find some documentation regarding how to tune the behavior of the spacy's lemmatizer (v2.2)
For example - Currently words like media and data get changed to medium/datum.
I'm interested in tweaking this to reflect how I want the lemmatizer to behave.
P.S - I am not interested in switching to a lookup based lemmatizer , I'm interested in tweaking the POS rule based lemmatizer behavior.
Could you please point me to any resources that explain how to do this?
Hi @thalish ,
The lemmatizer in v2.2 didn't really have a configurable API. In v3 the component is much better, although we're still working on end-to-end guides.
You should ask this type of question on the spaCy discussion forum, though, not here: Discussions · explosion/spaCy · GitHub
Oops! Just realized this is a pure spacy question and not a good fit for this forum.
I've gone ahead and created this thread , looking forward to learning more about spacy3.0 Custom spacy Lemmatizer · Discussion #7102 · explosion/spaCy · GitHub