metadata
dataset_info:
features:
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: label
dtype: int32
splits:
- name: train
num_bytes: 5850744
num_examples: 50484
download_size: 3399518
dataset_size: 5850744
configs:
- config_name: default
data_files:
- split: train
path: data/train-*