metadata
dataset_info:
features:
- name: input_ids
sequence: int32
splits:
- name: train
num_bytes: 39472739208
num_examples: 4816098
download_size: 14183439117
dataset_size: 39472739208
configs:
- config_name: default
data_files:
- split: train
path: data/train-*