NER for German Legal Text in Flair (default model) This is the legal NER model for German that ships with Flair. F1 Score: 96,35 (LER German dataset) Predicts 19 tags: tag meaning AN Anwalt EUN Europäische Norm GS Gesetz GRT Gericht INN Institution LD Land LDS Landschaft LIT Literatur MRK Marke ORG Organisation PER Person RR Richter RS Rechtssprechung ST Stadt STR Straße UN Unternehmen VO Verordnung VS Vorschrift VT Vertrag Based on Flair embeddings and LSTM CRF. More details on the Legal NER dataset here Demo: How to use in Flair Requires: Flair ( pip install flair ) This yields the following output: So, the entities " W. " (labeled as a person ) and " § 36 Abs. 7 IfSG " (labeled as a Gesetz ) are found in the sentence " Herr W. verstieß gegen § 36 Abs. 7 IfSG. ". Training: Script to train this model The following Flair script was used to train this model: Cite Please cite the following papers when using this model. Issues? The Flair issue tracker is available here.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy