metadata
license: mit
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: answer_start
dtype: int64
splits:
- name: train
num_bytes: 7163998
num_examples: 2617
download_size: 648021
dataset_size: 7163998
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Dataset Card for Dataset Name
This is dataset is ported to huggingface directly from the semantic-systems/amharic-qa repo. Please read the paper in the provided link
- Curated by: Tilahun Abedissa, Ricardo Usbeck, Yaregal Assabie
- Shared by: Dagim Ashenafi
- Language(s) (NLP): Amharic
- License: MIT
Dataset Sources
- Repository: https://github.com/semantic-systems/amharic-qa
- Paper: https://arxiv.org/abs/2303.03290
Dataset Structure
This dataset contains information in a tabular format with three columns ('question, 'answer', 'context') and a total of 2617 rows. Each row represents a data example with a question and its corresponding answer.
Source Data
Arrticles collected from Amharic Wikipedia dump file.
Data Collection and Processing
Crowdsourcing and annotated using the Haystack QA annotation tool.
Citation
Abedissa, T., Usbeck, R., & Assabie, Y. (Year). AmQA: Amharic Question Answering Dataset. Addis Ababa University, Universität Hamburg. https://orcid.org/0000-0002-0191-7211
Dataset Card Authors
Dagim Ashenafi