README.md

---
license: mit
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
  - split: valid
    path: data/valid-*
dataset_info:
  features:
  - name: id
    dtype: string
  - name: ner_tags
    sequence:
      class_label:
        names:
          '0': O
          '1': B-PERS
          '2': I-PERS
          '3': B-ORG
          '4': I-ORG
          '5': B-LOC
          '6': I-LOC
          '7': B-PROD
          '8': I-PROD
          '9': B-DATE
          '10': I-DATE
  - name: tokens
    sequence: string
  - name: doc_id
    dtype: string
  - name: region
    dtype: string
  splits:
  - name: train
    num_bytes: 4446694.039497307
    num_examples: 8020
  - name: test
    num_bytes: 556113.9802513465
    num_examples: 1003
  - name: valid
    num_bytes: 556113.9802513465
    num_examples: 1003
  download_size: 1458677
  dataset_size: 5558922
task_categories:
- token-classification
language:
- ro
tags:
- ner
- history
- romanian
size_categories:
- 10K<n<100K
---

# HistNERo

This repository contains the Historical Romanian Named Entity Recognition (HistNERo) dataset. The dataset has 10,026 sentences (i.e. 8,020 train, 1,003 valid and 1,003 train) annotated with five named entities: PERSON, ORGANIZATION, LOCATION, PRODUCT and DATE.
You can read more about the dataset in this [*HistNERo: Historical Named Entity Recognition for the Romanian Language*](https://arxiv.org/abs/2405.00155)

## Loading

To load the dataset, simply run:

```Python
from datasets import load_dataset

dataset = load_dataset("avramandrei/histnero")
```

Each sample respects the following format:
```json
{
  "id": "528",
  "ner_tags": [0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0, 0, 3, 4, 4, 4, 4, 0],
  "tokens": ["maĭ", "incóce", "vidu", "locuitoriĭ", "suburbiilorŭ", ",", "iar", "la", "riul", "Sabiĭului", "de", "càtrâ", "bisericâ", "romanéscâ", "gr.", "unitá", "Mai", "."],
  "doc_id": "Brasov_20-_20Gazeta_20Transilvaniei_201852.ann",
  "region": "Transylvania"
}
``` 

## Citation

Work in progress...