Deep-NER: named entity recognizer based on ELMo or BERT as embeddings and CRF as final classifier
Project description
Named entity recognizer based on ELMo or BERT as feature extractor and CRF as final classifier.
The goal of this project is creation of a simple Python package with the sklearn-like interface for solution of different named entity recognition tasks in case number of labeled texts is very small (not greater than several thousands). Special neural network language models named as ELMo (Embeddings from Language Models) and BERT (Bidirectional Encoder Representations from Transformers) ensure this possibility, because these language model were pre-trained on large text corpora and so they can select deep semantic features from text.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.