README.md

metadata

language:
  - da
license: apache-2.0
tags:
  - token-classification
  - ner
  - danish
  - modernbert
datasets:
  - alexandrainst/dane
metrics:
  - f1
  - precision
  - recall
pipeline_tag: token-classification
widget:
  - text: Jens Peter Hansen bor i København og arbejder hos Novo Nordisk.

ModernBERT Danish NER (Base)

Danish Named Entity Recognition model fine-tuned from AI-Sweden-Models/ModernBERT-base on the DaNE dataset.

Benchmark: DaNE Test Set

Entity	Precision	Recall	F1	Support
PER	0.8962	0.9061	0.9011	181
ORG	0.6929	0.6299	0.6599	154
LOC	0.7500	0.8969	0.8169	97
MISC	0.4878	0.6316	0.5505	95
micro avg	0.7260	0.7742	0.7493

Entity Types

• PER: Person names
• ORG: Organizations
• LOC: Locations
• MISC: Miscellaneous entities

Usage

from transformers import pipeline

ner = pipeline("ner", model="thomasbeste/modernbert-da-ner-base", aggregation_strategy="simple")
results = ner("Jens Peter Hansen bor i København og arbejder hos Novo Nordisk.")
for entity in results:
    print(f"{entity['word']}: {entity['entity_group']} ({entity['score']:.3f})")

Training Details

• Base model: AI-Sweden-Models/ModernBERT-base
• Dataset: DaNE (alexandrainst/dane) — 4,383 train / 564 val / 565 test sentences
• Epochs: 10
• Learning rate: 2e-5
• Batch size: 16
• Optimizer: AdamW (weight decay 0.01, warmup ratio 0.1)
• Precision: bf16
• Max sequence length: 256