metadata
dataset_info:
features:
- name: input_ids
list: int32
- name: attention_mask
list: int8
- name: labels
list: int64
splits:
- name: train
num_bytes: 4589328582
num_examples: 5222418
- name: test
num_bytes: 33766836
num_examples: 37519
- name: valid
num_bytes: 43685664
num_examples: 51809
download_size: 5732466191
dataset_size: 4666781082
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*