metadata
dataset_info:
features:
- name: query
dtype: string
- name: positive
dtype: string
- name: negative
dtype: string
splits:
- name: train
num_bytes: 48101906754
num_examples: 26111683
download_size: 6845539261
dataset_size: 48101906754
configs:
- config_name: default
data_files:
- split: train
path: data/train-*