metadata
dataset_info:
features:
- name: SeqA
dtype: string
- name: SeqB
dtype: string
- name: labels
dtype: int64
splits:
- name: train
num_bytes: 781699363
num_examples: 679827
- name: valid
num_bytes: 2828661
num_examples: 2500
- name: test
num_bytes: 2975087
num_examples: 2500
download_size: 729420086
dataset_size: 787503111
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*
Human PiNUI set with 2500 random validation and test examples