metadata
dataset_info:
features:
- name: query
dtype: string
- name: positives
sequence: string
- name: negatives
sequence: string
splits:
- name: train
num_bytes: 850663698
num_examples: 50000
download_size: 494705376
dataset_size: 850663698
configs:
- config_name: default
data_files:
- split: train
path: data/train-*