BERT 101 - State Of The Art NLP Model Explained
Hugging Face Blog
Google researchers developed BERT in 2018, a machine learning model that handles over 11 common natural language processing tasks like sentiment analysis and named entity recognition using a bidirectional learning approach. BERT was trained on 3.3 billion words from Wikipedia and Google's BooksCorpus using 64 tensor processing units over 4 days, and Google integrated it into search in November 2020 to better understand query context. The model's open-source availability allows developers to fine-tune it for specific applications without building separate models for each language task.
