metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
splits:
- name: train
num_bytes: 1938611023.370665
num_examples: 494347
- name: validation
num_bytes: 911999.34
num_examples: 238
download_size: 1120050006
dataset_size: 1939523022.710665
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*