Datasets:
metadata
language:
- ja
dataset_info:
features:
- name: passage_id
dtype: int64
- name: title
dtype: string
- name: section_title
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 6026524
num_examples: 6930
download_size: 3331615
dataset_size: 6026524
configs:
- config_name: default
data_files:
- split: train
path: data/train-*