metadata
dataset_info:
features:
- name: input
dtype: string
- name: type
dtype: string
- name: outputs
list:
- name: answer
dtype: string
- name: question
dtype: string
- name: summary
dtype: string
- name: document_id
dtype: string
splits:
- name: train
num_bytes: 2896878878
num_examples: 140866
download_size: 1159537158
dataset_size: 2896878878
configs:
- config_name: default
data_files:
- split: train
path: data/train-*