Closed gloignon closed 5 years ago
interesting :) I wonder what lemmatizer they used.
The examples you give do not come from the corpus:
I guess that your examples are outputs of the UDPIPE model trained on UD 2.3.
There are not enough data in the UD corpus to "learn" the lemma cuire correctly, so I guess that the parser use the most productive verbs of French "1er groupe" and predict by a kind of analogy the lemma cuiser
Dans UD2.3, certaines conjugaisons du verbe "cuire" sont lemmatisées comme étant le verbe "cuiser" plutôt que "cuire". Par exemple:
With UD2.3, some tenses of the verb "cuire" are lemmatized as "cuiser". For example:
Output: