README.md

---
language:
  - da
license: mit
tags:
  - named-entity-recognition
  - token-classification
  - danish
  - ner
  - xlm-roberta
  - pii-detection
  - gdpr
datasets:
  - chcaa/dansk-ner
  - KennethEnevoldsen/dane_plus
  - ltg/norne
pipeline_tag: token-classification
model-index:
  - name: danish-xlmr-ner-large
    results:
      - task:
          type: token-classification
          name: Named Entity Recognition
        dataset:
          type: chcaa/dansk-ner
          name: DANSK (dev)
        metrics:
          - type: f1
            value: 87.6
            name: F1 (micro)
          - type: precision
            value: 86.6
            name: Precision
          - type: recall
            value: 88.7
            name: Recall
      - task:
          type: token-classification
          name: Named Entity Recognition
        dataset:
          type: chcaa/dansk-ner
          name: DANSK (test)
        metrics:
          - type: f1
            value: 86.5
            name: F1 (micro)
          - type: precision
            value: 85.4
            name: Precision
          - type: recall
            value: 87.5
            name: Recall
      - task:
          type: token-classification
          name: Named Entity Recognition
        dataset:
          type: KennethEnevoldsen/dane_plus
          name: DaNE (dev)
        metrics:
          - type: f1
            value: 93.0
            name: F1 (micro)
          - type: precision
            value: 93.2
            name: Precision
          - type: recall
            value: 92.9
            name: Recall
      - task:
          type: token-classification
          name: Named Entity Recognition
        dataset:
          type: KennethEnevoldsen/dane_plus
          name: DaNE (test)
        metrics:
          - type: f1
            value: 87.7
            name: F1 (micro)
          - type: precision
            value: 88.1
            name: Precision
          - type: recall
            value: 87.3
            name: Recall
---

# Danish XLM-R NER Large (Two-Stage)

A Danish Named Entity Recognition model based on `xlm-roberta-large` (560M parameters), fine-tuned in two stages for high-recall PII detection in Danish text.

## Model Description

This model detects three entity types relevant for GDPR-compliant PII processing:

- **PER** - Person names
- **ORG** - Organizations (companies, institutions, government bodies)
- **LOC** - Locations (addresses, cities, countries)

MISC is intentionally excluded to reduce noise and focus on actionable PII entities.

### Training

**Two-stage fine-tuning approach:**

1. **Stage 1** (broad NER): DANSK + DaNE + NorNE, 10 epochs, LR 2e-5
2. **Stage 2** (domain adaptation): DANSK-only, 1 epoch, LR 5e-6

This approach achieves the best balance between multi-domain generalization and Danish-specific performance.

### Datasets

| Dataset | Role | Size | Domains |
|---------|------|------|---------|
| [DANSK](https://huggingface.co/datasets/chcaa/dansk-ner) | Primary | 11.7K train | Web, News, Wiki, Legal, Dannet, Conversation, Social Media |
| [DaNE](https://huggingface.co/datasets/KennethEnevoldsen/dane_plus) | Supplementary | 4.4K train | News |
| [NorNE](https://huggingface.co/datasets/ltg/norne) | Stage 1 only | ~20K train | News (Norwegian Bokmaal + Nynorsk) |

## Evaluation Results

### DANSK (primary benchmark, multi-domain)

| Split | PER F1 | ORG F1 | LOC F1 | Micro F1 | Precision | Recall |
|-------|--------|--------|--------|----------|-----------|--------|
| Dev   | 88.0   | 85.3   | 90.3   | **87.6** | 86.6      | 88.7   |
| Test  | 84.8   | 84.6   | 90.3   | **86.5** | 85.4      | 87.5   |

### DaNE (secondary benchmark, news domain)

| Split | PER F1 | ORG F1 | LOC F1 | Micro F1 | Precision | Recall |
|-------|--------|--------|--------|----------|-----------|--------|
| Dev   | 97.5   | 85.1   | 92.9   | **93.0** | 93.2      | 92.9   |
| Test  | 94.2   | 79.7   | 87.8   | **87.7** | 88.1      | 87.3   |

### GPI Legal Documents (independent evaluation, Danish legal domain)

Evaluated on 30 human-corrected documents (contracts, invoices, case briefs, client letters):

| Entity | Precision | Recall | Notes |
|--------|-----------|--------|-------|
| PER    | 0.76      | 1.00   | Perfect recall; FPs are email addresses misclassified as PER |
| ORG    | 0.94      | 0.96   | Near-perfect |
| LOC    | 0.52      | 0.51   | Boundary errors (detects street, misses house number) — detection rate is near-perfect |

LOC score reflects strict span matching. The model consistently detects location entities but predicts shorter spans (e.g., "Gothersgade" instead of "Gothersgade 81"). A post-processing step to extend LOC spans to include adjacent numbers resolves this.

## Usage

```python
from transformers import AutoModelForTokenClassification, AutoTokenizer, pipeline

model_name = "thomasbeste/danish-xlmr-ner-large"
nlp = pipeline("ner", model=model_name, aggregation_strategy="simple")

text = "Anders Jensen fra Danske Bank bor på Vestergade 42 i København."
entities = nlp(text)

for ent in entities:
    print(f"  {ent['entity_group']}: {ent['word']} (score: {ent['score']:.3f})")
```

### ONNX Deployment

For production use, export to ONNX INT8 for ~3x CPU speedup:

```bash
pip install optimum[onnxruntime]

# Export to ONNX
optimum-cli export onnx --model thomasbeste/danish-xlmr-ner-large ./model-onnx --task token-classification

# Quantize to INT8
python -c "
from optimum.onnxruntime import ORTQuantizer
from optimum.onnxruntime.configuration import AutoQuantizationConfig
q = ORTQuantizer.from_pretrained('./model-onnx')
qconfig = AutoQuantizationConfig.avx512_vnni(is_static=False, per_channel=False)
q.quantize(save_dir='./model-onnx-int8', quantization_config=qconfig)
"
```

## Label Scheme

IOB2 format with 7 labels:

| ID | Label |
|----|-------|
| 0  | O     |
| 1  | B-PER |
| 2  | I-PER |
| 3  | B-ORG |
| 4  | I-ORG |
| 5  | B-LOC |
| 6  | I-LOC |

## Intended Use

Designed for GDPR-compliant PII detection in Danish enterprise document processing pipelines. Optimized for **recall over precision** — a missed entity (false negative) is a compliance risk, while over-detection (false positive) is safe.

## Limitations

- Optimized for Danish text. May work on other Scandinavian languages (Norwegian, Swedish) but not evaluated.
- LOC boundary detection tends to predict shorter spans than the full address. Post-processing recommended.
- Email addresses are sometimes misclassified as PER. Downstream validation (reject names containing `@`) is recommended.