metadata
dataset_info:
features:
- name: passage_id
dtype: string
- name: passage
dtype: string
- name: type
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: short_answer
dtype: string
- name: short_answer_variants
list: 'null'
- name: source
dtype: string
- name: loss_extra_text
dtype: string
- name: inner_docs
list: string
- name: inner_doc_answers
list: string
splits:
- name: train
num_bytes: 20962312
num_examples: 4416
- name: validation
num_bytes: 381549
num_examples: 95
- name: test
num_bytes: 333414
num_examples: 91
download_size: 4007624
dataset_size: 21677275
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*