metadata
dataset_info:
features:
- name: anchor
dtype: string
- name: pos_neg
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 674717980
num_examples: 1095846
download_size: 160768607
dataset_size: 674717980
configs:
- config_name: default
data_files:
- split: train
path: data/train-*