Closed omarabb315 closed 1 month ago
Are you referring to the BERTUnfactoredDisambiguator?
Unfortunately there's no way to do this within CAMeL Tools at the moment (although that may change in the future).
Furthermore, it's important to note that our model isn't fine-tuned on lemmas but rather we produce lemmas as a byproduct of predicting other features.
You can read more in this paper. We also provide the original CAMeLBERT models (GitHub, Hugging Face) which you can use.
Is there any way to finetune the lemmatization tool?