metadata
dataset_info:
features:
- name: anchor
dtype: string
- name: positive
dtype: string
splits:
- name: train
num_bytes: 2644044179
num_examples: 32249997
- name: test
num_bytes: 139185172
num_examples: 1697369
download_size: 1726061402
dataset_size: 2783229351
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*