metadata
dataset_info:
features:
- name: premises
dtype: string
- name: hypothesis
dtype: string
- name: labels
dtype: int64
splits:
- name: train
num_bytes: 104700737
num_examples: 677489
download_size: 20846026
dataset_size: 104700737
configs:
- config_name: default
data_files:
- split: train
path: data/train-*