chunked/nltk_chunking/model_doc_distilbert/chunk_16.txt

Trained by distillation of the pretrained BERT model, meaning it’s been trained to predict the same probabilities as the larger model.