metadata
dataset_info:
features:
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: label
dtype:
class_label:
names:
'0': not_entailment
'1': entailment
splits:
- name: train
num_bytes: 5477208
num_examples: 20369
- name: test
num_bytes: 900112
num_examples: 3350
download_size: 3749404
dataset_size: 6377320
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*