metadata
dataset_info:
features:
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: label
dtype:
class_label:
names:
'0': entailment
'1': neutral
'2': contradiction
splits:
- name: train
num_bytes: 136716914
num_examples: 1912390
download_size: 48712225
dataset_size: 136716914
configs:
- config_name: default
data_files:
- split: train
path: data/train-*