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.

By filling the form below, you understand that only the metadata and the annotations of this corpus (sPBC) have a cc0-1.0 license, and that the rest of the content have each different licenses for which the authors of this corpus do not hold any copyright whatsoever (see the license). Instead of redistributing the data, you can always link to it here. We maintain the data by preserving previous versions, adding new languages in newer versions, and correcting any mistakes in the metadata. Noted that some parts of the data from certain sources cannot be redistributed publicly and are provided here solely for research purposes. Requests are manually reviewed, and access grants take 1-3 days. We are not obliged to provide access to all who fill the form. The sources of all data are open (see the license), and anyone can obtain them with some effort.

sPBC-1401-103

This is a super-parallel Bible corpus containing 1401 language labels (language_script pairs), meaning that for each verse, we have the translation in other languages. However, this results in a relatively low number of verses, with the current version supporting 103 verses per language.

Format

You can download all languages in one zip file (https://huggingface.co./datasets/cis-lmu/sPBC/resolve/main/103.zip). Each language pair is a text file with 103 lines, and the lines are parallel and in order.

License and Citation

For the citation and proper usage, see the license (https://huggingface.co./datasets/cis-lmu/sPBC/blob/main/LICENSE).

This dataset is used to compute MEXA scores in the following paper: https://arxiv.org/abs/2410.05873