chunked/nltk_chunking/model_doc_distilbert/chunk_7.txt

Our smaller, faster and lighter model is cheaper to pre-train and we
demonstrate its capabilities for on-device computations in a proof-of-concept experiment and a comparative on-device
study.