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: 65884386
num_examples: 550152
- name: validation
num_bytes: 1263036
num_examples: 10000
- name: test
num_bytes: 1258904
num_examples: 10000
download_size: 20183884
dataset_size: 68406326
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*