metadata
dataset_info:
features:
- name: anchor
dtype: string
- name: positive
dtype: string
- name: negative
dtype: string
splits:
- name: train
num_bytes: 24930485
num_examples: 40000
download_size: 14294145
dataset_size: 24930485
configs:
- config_name: default
data_files:
- split: train
path: data/train-*