metadata
dataset_info:
features:
- name: id
dtype: string
- name: title
dtype: string
- name: context
dtype: string
- name: question
dtype: string
- name: answers
struct:
- name: answer_start
sequence: int64
- name: text
sequence: string
- name: masked_query
dtype: string
- name: query_embedding
sequence: float64
splits:
- name: train
num_bytes: 697771735
num_examples: 97888
download_size: 506779114
dataset_size: 697771735
configs:
- config_name: default
data_files:
- split: train
path: data/train-*