metadata
dataset_info:
features:
- name: sentence1
dtype: string
- name: sentence2
dtype: string
- name: score
dtype: float64
- name: similarity
dtype: float64
- name: type
dtype: string
- name: hard
dtype: int64
- name: label
dtype: float64
splits:
- name: train
num_bytes: 31138579811
num_examples: 11052543
- name: test
num_bytes: 113404662
num_examples: 40000
- name: eval
num_bytes: 28643514
num_examples: 10000
download_size: 17826299042
dataset_size: 31280627987
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: eval
path: data/eval-*