Commit History
-
Train Punkt model on Vietnamese data for sentence segmentation
- Add train_punkt.py: fetch data from 4 HF datasets and train NLTK PunktTrainer - Add punkt_params_trained.json: 672 abbrevs, 378 starters, 3264 collocations - Add sent_tokenize.py: tokenizer using trained Punkt model - Update evaluate.py: add --improved flag for trained model evaluation - Add underthesea submodule9ec538b
