flair/ner-german

This is the standard 4-class NER model for German that ships with Flair . The F1-Score: 87,94 (CoNLL-03 German revised)Predicts 4 tags: "George Washington" and "Washington" The model is based on Flair embeddings and LSTM-CRF . The following Flair script was used to train this model: How to use in Flair (pip install flair) and how to use the model: Flair, GloVe, StackedEmbeddings, FlairEmbedding, SequenceTagger . The code is available to download and use for the next Flair version of the F1,