README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: string
    - name: context
      dtype: string
    - name: question
      dtype: string
    - name: answers
      struct:
        - name: answer_start
          list: int64
        - name: text
          list: string
    - name: lang
      dtype: string
  splits:
    - name: train
      num_bytes: 17768236.09346378
      num_examples: 11529
    - name: validation
      num_bytes: 2220836.864487927
      num_examples: 1441
    - name: test
      num_bytes: 2222378.042048293
      num_examples: 1442
  download_size: 13871572
  dataset_size: 22211451
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
license: cc-by-4.0
task_categories:
  - question-answering
language:
  - sw
  - en
  - luo
  - lug
  - luy
  - kik
  - yor
  - kln
  - kin
  - lin
  - mas
  - hau
  - zul
  - xho
  - twi
  - fon
  - guz
tags:
  - african-languages
  - QuAD
  - Cross-lingual
size_categories:
  - 1K<n<10K

AfriQuAD (Multilingual)

This dataset spans across different African languages and also touches on cross-lingual aspects. It can be best used in Question Answering tasks.

from datasets import load_datasets

#load the entire dataset
dataset = load_dataset("theophilusowiti/AfriQuAD")