metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 298066739240
num_examples: 32000000
- name: test
num_bytes: 4314590
num_examples: 1000
download_size: 115876227740
dataset_size: 298071053830
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*