metadata
dataset_info:
features:
- name: anchor
dtype: string
- name: positive
dtype: string
- name: negative
dtype: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 1281750763
num_examples: 3401694
download_size: 811108906
dataset_size: 1281750763
configs:
- config_name: default
data_files:
- split: train
path: data/train-*