metadata
dataset_info:
features:
- name: promptID
dtype: int32
- name: pairID
dtype: string
- name: premise
dtype: string
- name: premise_binary_parse
dtype: string
- name: premise_parse
dtype: string
- name: hypothesis
dtype: string
- name: hypothesis_binary_parse
dtype: string
- name: hypothesis_parse
dtype: string
- name: genre
dtype: string
- name: label
dtype:
class_label:
names:
'0': entailment
'1': neutral
'2': contradiction
- name: premise_tokens
sequence: int64
- name: hypothesis_tokens
sequence: int64
- name: full_tokens
sequence: int64
- name: G
dtype: int64
- name: 'Y'
dtype: bool
splits:
- name: train
num_bytes: 37026836.81773966
num_examples: 4000
- name: validation_matched
num_bytes: 90665914
num_examples: 9815
download_size: 8880876
dataset_size: 127692750.81773967
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation_matched
path: data/validation_matched-*