README.md

---
license: apache-2.0
task_categories:
- translation
language:
- si
tags:
- transliteration
size_categories:
- 1M<n<10M
---
# Sinhala Romanized Dataset

This dataset contains Sinhala text with romanized (transliterated) versions, created using publicly available Sinhala data sources.

## Dataset Description

The Augmented Sinhala to Romanized Sinhala Dataset provides paired examples of Sinhala text and their corresponding romanized transliterations. This dataset aims to facilitate research in Sinhala language processing, particularly for applications that require romanized representations of Sinhala text.

## Sources

The data was compiled from the following publicly available datasets:
- [SemiSOLD](https://huggingface.co/datasets/sinhala-nlp/SemiSOLD)
- [Sinhala Sentiment Analysis Dataset](https://huggingface.co/datasets/sinhala-nlp/sinhala-sentiment-analysis)
- [NSINA](https://huggingface.co/datasets/sinhala-nlp/NSINA)

## Transliteration

The romanized (transliterated) version of the Sinhala text was generated using a standardized transliteration process. 
The transliteration follows standard patterns for converting Sinhala characters to Roman script.

**Note**: This dataset uses standardized transliteration patterns and does not capture ad hoc transliterations 
that might be used in informal contexts.

## Usage

```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("deshanksuman/Augmented_SinhalatoRomanizedSinhala_Dataset")

# Example usage
for example in dataset['train'][:5]:
    print(f"Sinhala: {example['sinhala_text']}")
    print(f"Romanized: {example['romanized_text']}")
    print("---")
```

## Citation

If you use this dataset in your research, please cite:

```bibtex
@misc{suman2025augmentedsinhala,
  author = {Deshan Sumanathilaka},
  title = {Augmented Sinhala to Romanized Sinhala Dataset},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/deshanksuman/Augmented_SinhalatoRomanizedSinhala_Dataset}}  
}
```

Please also cite the original sources as appropriate:

- SemiSOLD dataset:
```bibtex
@article{ranasinghe2022sold,
  title={SOLD: Sinhala Offensive Language Dataset},
  author={Ranasinghe, Tharindu and Anuradha, Isuri and Premasiri, Damith and Silva, Kanishka and Hettiarachchi, Hansi and Uyangodage, Lasitha and Zampieri, Marcos},
  journal={arXiv preprint arXiv:2212.00851},
  year={2022}
}
```

- NSINA dataset:
```bibtex
@inproceedings{Nsina2024,
author={Hettiarachchi, Hansi and Premasiri, Damith and Uyangodage, Lasitha and Ranasinghe, Tharindu},
title={{NSINA: A News Corpus for Sinhala}},
booktitle={The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},
year={2024},
month={May},
}
```

## License

This dataset is released under the same licenses as the original source datasets. Please refer to the original datasets for specific licensing information.

## Acknowledgements

We thank the creators of the original datasets for making their data publicly available.