metadata
dataset_info:
features:
- name: input_ids
sequence: uint16
splits:
- name: train
num_bytes: 28156723200
num_examples: 13721600
download_size: 27952802456
dataset_size: 28156723200
configs:
- config_name: default
data_files:
- split: train
path: data/train-*