metadata
dataset_info:
features:
- name: query
dtype: string
- name: positive
sequence: string
- name: negative
sequence: string
splits:
- name: train
num_bytes: 122780745
num_examples: 18906
download_size: 11992524
dataset_size: 122780745
configs:
- config_name: default
data_files:
- split: train
path: data/train-*