metadata
dataset_info:
features:
- name: anchor
dtype: large_string
- name: positive
dtype: large_string
- name: class
dtype: large_string
splits:
- name: train
num_bytes: 9753922
num_examples: 40906
- name: val
num_bytes: 1621859
num_examples: 6585
download_size: 1884863
dataset_size: 11375781
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*