chunked/nltk_chunking/_tasks_explained/chunk_133.txt

Text classification
BERT is an encoder-only model and is the first model to effectively implement deep bidirectionality to learn richer representations of the text by attending to words on both sides.