metadata
language:
- en
dataset_info:
features:
- name: article
dtype: string
- name: article_sha1
dtype: string
- name: section_idx
dtype: int32
- name: section_title
dtype: string
- name: content
dtype: string
- name: question
dtype: string
- name: is_adversarial
dtype: bool
- name: input_ids
sequence: int32
- name: token_type_ids
sequence: int8
- name: attention_mask
sequence: int8
- name: label
dtype: int64
- name: predictions-00
dtype: int64
- name: predictions-01
dtype: int64
- name: predictions-02
dtype: int64
- name: predictions-03
dtype: int64
- name: predictions-04
dtype: int64
splits:
- name: train
num_bytes: 4070875769
num_examples: 706498
- name: validation
num_bytes: 85304716
num_examples: 15000
download_size: 1563522291
dataset_size: 4156180485
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*