Bapynshngain/Bapyn-EnKha-Parallel-Corpus

This dataset is archived on Zenodo:

DOI: https://doi.org/10.5281/zenodo.18703428

Note: This dataset will be readily granted access upon request. The form is only to log users and make sure they adhere to the license terms and conditions

To request access:

1. Submit an access request via the Hugging Face dataset page.
2. Complete the following access form: 👉 Request Access Form

Please note that requests will be reviewed and verified by matching your Hugging Face username and details with the information provided in the form. Access will be granted only after successful verification.

This is a High-Quality Parallel Corpus personally curated by Bapynshngainlang Nongkynrih.

• It contains a total of 40,000 pairs of English-Khasi pairs
• 26,000 pairs was taken from IndicNECorp1.0 from the paper "Findings of the WMT 2023 Shared Task on Low-Resource Indic Language Translation" by Santanu Pal, Partha Pakray, Sahinur Rahman Laskar, Lenin Laitonjam, Vanlalmuansangi Khenglawt, Sunita Warjri, Pankaj Kundan Dadure and Sandeep Kumar Dash
• 1,425 taken from Tatoeba project English-Khasi dataset.
• Rest of the dataset is sourced from publicly availaible Khasi resources which were then manually translated by me.
• This is an upgrade over my old dataset, namely, Bapynshngain/English-Khasi-Parallel-Corpus.