metadata
dataset_info:
features:
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: label
dtype: int64
- name: source
dtype: string
splits:
- name: train
num_bytes: 30134573
num_examples: 105000
download_size: 18361633
dataset_size: 30134573
configs:
- config_name: default
data_files:
- split: train
path: data/train-*