metadata
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 34518014532
num_examples: 66895377
download_size: 21272361921
dataset_size: 34518014532
configs:
- config_name: default
data_files:
- split: train
path: data/train-*