Datasets:

dalaone
/

dafand

configs:
  - config_name: edo_english_bible
    data_files:
      - split: train
        path: edo_english_bible/train.json
  - config_name: edo_english_tales
    data_files:
      - split: train
        path: edo_english_tales/train.json
  - config_name: edo_jw_lessons
    data_files:
      - split: train
        path: edo_jw_lessons/train.json
license: apache-2.0
task_categories:
  - translation
language:
  - en
  - edo
pretty_name: a
size_categories:
  - 10K<n<100K

Dataset Description

A parallel corpus for [language pair] translation, following the MAFAND dataset format by Masakhane.

Usage

import datasets

data = datasets.load_datasets("dalaone/dafand", "edo_english_bible")

Data

edo_english_bible
edo_english_tales
edo_jw_lessons

Dataset Structure

{
  "translation": {
    "edo": "source sentence",
    "english": "target sentence"
  }
}

Citation

Inspired by / adapted from Masakhane's github and the MAFAND dataset hugging-face

Limitations

Bible-biased due to limited training data