wpnbos's picture
Update README.md
13d1798
metadata
language:
  - nl
tags:
  - Named Entity Recognition
  - xlm-roberta
datasets:
  - conll2002
metrics:
  - f1: 90.57

XLM-RoBERTa base ConLL-2002 Dutch

XLM-Roberta base model finetuned on ConLL-2002 Dutch train set, which is a Named Entity Recognition dataset containing the following classes: PER, LOC, ORG and MISC.

Results from https://arxiv.org/pdf/1911.02116.pdf reciprocated (original results were 90.39 F1, this finetuned version here scored 90.57).