metadata
dataset_info:
features:
- name: document_id
dtype: int32
- name: question
dtype: string
- name: id
dtype: int32
- name: answers
sequence: string
- name: chunks
sequence: string
- name: correct_chunks
sequence: int64
- name: num_words
dtype: int64
splits:
- name: train
num_bytes: 65392614
num_examples: 2019
download_size: 9307402
dataset_size: 65392614
configs:
- config_name: default
data_files:
- split: train
path: data/train-*