metadata
dataset_info:
features:
- name: query
dtype: string
- name: positive
dtype: string
- name: negative
dtype: string
- name: negatives
sequence: string
splits:
- name: train
num_bytes: 27593074
num_examples: 100000
download_size: 17493110
dataset_size: 27593074
configs:
- config_name: default
data_files:
- split: train
path: data/train-*