You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

๐Ÿง  Meitei Mayek Translation Dataset

A high-quality parallel corpus for English โ†” Meitei Mayek (Manipuri), designed to support low-resource NLP research, especially for translation, text generation, and TTS/STT applications. This dataset directly addresses the gap in digital resources for Meitei Lon written in Meitei Mayek script.


๐ŸŒ Language Overview

  • Language: Manipuri (Meitei Lon)
  • Script: Meitei Mayek
  • ISO 639-3: mni
  • Script Code: Mtei
  • Language Code: mni_Mtei
  • Region: Manipur, India ๐Ÿ‡ฎ๐Ÿ‡ณ
  • Type: Low-resource, underrepresented
  • Status: Official language of Manipur

๐Ÿ“ฆ Dataset Overview

Split Examples Size (bytes)
train 415,210 114.99 MB
validation 23,067 6.39 MB
test 23,068 6.39 MB
Total 461,345 127.77 MB
  • Download Size: 61.06 MB
  • Format: TSV (Tab-Separated Values)
  • License: MIT
  • Tokenizer Ready: โœ…
  • Text Normalized: โœ…

๐Ÿ“ Data Schema

Column Description
src_lang Source language code (e.g., eng_Latn)
tgt_lang Target language code (e.g., mni_Mtei)
english English sentence (source/reference)
meitei_mayek Translated sentence in Meitei Mayek script

๐Ÿ” Example

{
  "src_lang": "eng_Latn",
  "tgt_lang": "mni_Mtei",
  "english": "i am hungry",
  "meitei_mayek": "๊ฏ‘๊ฏฉ ๊ฏ†๊ฏฅ๊ฏ› ๊ฏ‚๊ฏฅ๊ฏ๊ฏƒ๊ฏฆ"
}

LOAD

from datasets import load_dataset

dataset = load_dataset("DayanandaThokchom/english-TO-meitei-mayek")

print(dataset["train"][0])

๐Ÿ“œ License

MIT License Free to use, modify, and distribute. Attribution appreciated!

๐Ÿ“ฌ Contact

This project is maintained by Dayananda Thokchom

Conclusion

Language is more than communicationโ€”it's identity, culture, and legacy. In the world of AI, languages like Meitei Mayek deserve not just preservation, but participation. With this dataset, weโ€™ve laid the foundation for equitable AI systems that respect and empower underrepresented languages.

Weโ€™re not just building a datasetโ€”weโ€™re building a future where Meitei Mayek can speak, listen, translate, and be understood by machines on the global stage.

Every byte of data brings Meitei Mayek closer to digital equality. Letโ€™s keep pushing boundaries, one token at a time.

Join the mission. Make your contribution count. ๐Ÿ’ช๐ŸŒ

Downloads last month
33