metadata
dataset_info:
features:
- name: query
dtype: string
- name: positive
dtype: string
- name: negatives
sequence: string
- name: dataset
dtype: string
splits:
- name: train
num_bytes: 1769503708
num_examples: 3455203
download_size: 996832662
dataset_size: 1769503708
configs:
- config_name: default
data_files:
- split: train
path: data/train-*