metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation_matched
path: data/validation_matched-*
- split: validation_mismatched
path: data/validation_mismatched-*
dataset_info:
features:
- name: promptID
dtype: int32
- name: pairID
dtype: string
- name: premise
dtype: string
- name: hypothesis
dtype: string
- name: genre
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 81317900
num_examples: 392702
- name: validation_matched
num_bytes: 2010024
num_examples: 9815
- name: validation_mismatched
num_bytes: 2121266
num_examples: 9832
download_size: 56640779
dataset_size: 85449190
license: cc-by-4.0