metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 184550589140
num_examples: 6928615
- name: test
num_bytes: 20505644692
num_examples: 769847
download_size: 62121422076
dataset_size: 205056233832
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*