metadata
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 35342459200
num_examples: 8620112
download_size: 17366763850
dataset_size: 35342459200
configs:
- config_name: default
data_files:
- split: train
path: data/train-*