John Snow Labs Releases Spark NLP 4.0
Delivers improved speed and over 1,000 new models to the most-used NLP library in the enterprise .
Note: TDWI’s editors carefully choose vendor-issued press releases about new or upgraded products and services. We have edited and/or condensed this release to highlight key features but make no claims as to the accuracy of the vendor's statements.
John Snow Labs, developer of the Spark NLP library, announced the release of Spark NLP 4.0. With new question-answering annotators, major performance improvements, optimizations on new hardware platforms, and more than 1,000 state-of-the-art, pre-trained transformer models available in multiple languages, Spark NLP 4.0 is the company’s most significant release this year.
The introduction of question-answering annotators in Spark NLP enables the software to answer arbitrary natural-language questions based on a given document. The models provide both an answer and an explanation of where in the document the answer came from. Hundreds of pre-trained models are available out-of-the-box, enabling support for multiple languages, document types, and performance goals. Models are trained and fine-tuned so users can start using these applications immediately.
Improvements to the accuracy of key tasks, delivering new state-of-the-art accuracy for two popular tasks, are also featured in this release. One is named entity recognition (NER), for which Spark NLP 4.0 now provides the most accurate model on the popular CoNLL-2003 benchmark among open source NLP libraries. The second is co-reference resolution, using BERT-based span classification to outperform traditional approaches and libraries.
More information is available at https://www.johnsnowlabs.com.