metadata
dataset_info:
features:
- name: id
dtype: string
- name: title
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: answers
sequence:
- name: text
dtype: string
- name: answer_start
dtype: int32
splits:
- name: train
num_bytes: 116732025
num_examples: 130319
- name: validation
num_bytes: 5722017.701170724
num_examples: 5826
- name: evaluation
num_bytes: 196430.40512086247
num_examples: 200
- name: validation_bad
num_bytes: 5742642.893708414
num_examples: 5847
download_size: 21647341
dataset_size: 128393116
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: evaluation
path: data/evaluation-*
- split: validation_bad
path: data/validation_bad-*