Datasets:

DusunDictionary
/

dusun-dictionary-corpus

language:
  - dtp
  - en
  - ms
license: mit
pretty_name: Dusun-English-Malay Dictionary and Corpus
task_categories:
  - translation
tags:
  - conditional-text-generation
  - nmt
  - low-resource
  - endangered-language
  - dusun
size_categories:
  - 10K<n<100K

Dusun-English-Malay Dictionary and Corpus

A trilingual dataset containing Dusun, English, and Malay words, phrases, and sentences compiled for linguistic research, dictionary development, and machine translation. This dataset is based on the Dusun language as spoken by the Dusun ethnic group of Sabah, Malaysia. While the Dusun dialect in the dataset shares approximately 99% similarity with standardized Kadazandusun, there may be minor differences in vocabulary, spelling, and standardisation. Researchers working on Kadazandusun NLP may find this dataset largely applicable.

Dataset Overview

This dataset contains approximately 13,000 entries including:

~6k+ dictionary words with translations
~5K+ phrases and sentences with translations

Languages: Dusun (dtp), English (en), Malay (ms)

Curated from DusunDictionary.com for Neural Machine Translation (NMT) model training and language preservation.

Dataset Structure

Columns

Column	Description
Dusun	Word or phrase in Dusun language
English	English translation
Type	Part-of-speech or entry type (noun, verb, adj, phrase, etc.)
Malay	Malay translation
Category	Semantic/thematic category
Variant	Dialectal variations and synonyms

Example Entry

Dusun,English,Type,Malay,Category,Variant
osonong,good,adj,bagus,quality,
doho,mine,pronoun,saya punya,,dogo
louson oku, I'm hungry,phrase,saya lapar,,

Usage

Loading the Dataset

from datasets import load_dataset

# Load full dataset
dataset = load_dataset("DusunDictionary/dusun-dictionary-corpus")

# Access data
print(dataset['train'][0])

Filtering for NMT Training

Extract sentence-level data for machine translation:

# Filter phrases only (adjust based on your Type values)
phrases = dataset['train'].filter(lambda x: x['Type'] == 'phrase')

# Create Dusun-English pairs
dusun_english = phrases.map(lambda x: {
    'source': x['Dusun'],
    'target': x['English']
})

Example Use Cases

Machine Translation: Training Dusun↔English and Dusun↔Malay models
Language Preservation: Documenting endangered language vocabulary
Linguistic Research: Studying Austronesian language structures
Educational Tools: Building language learning applications

Language Information

Dusun (ISO 639-3: dtp) is an indigenous language spoken in Sabah, Malaysia. It is considered endangered, with fewer younger generation speakers. This dataset is part of ongoing efforts to preserve and revitalize the language through technology.

Data Collection

Data is continuously curated and verified by native speakers through DusunDictionary.com, combining traditional lexicography with community contributions.

License

MIT License - Free to use with attribution.

Citation

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

@misc{dusun-corpus-2026,
  title={Dusun-English-Malay Dictionary and Corpus},
  author={DusunDictionary.com},
  year={2026},
  url={https://huggingface.co/datasets/DusunDictionary/dusun-dictionary-corpus}
}

Contact & Contributing

Website: DusunDictionary.com
Issues: Report errors or suggest improvements via Hugging Face discussions
Contributing: Native speakers and linguists welcome to contribute

Acknowledgments

This project is dedicated to preserving the Dusun language for future generations.