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
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 15695054
num_examples: 3696
- name: validation
num_bytes: 1726883
num_examples: 440
download_size: 3596346
dataset_size: 17421937
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*