metadata
dataset_info:
features:
- name: input_ids
sequence: uint16
splits:
- name: train
num_bytes: 103801651200
num_examples: 50585600
download_size: 103057426049
dataset_size: 103801651200
configs:
- config_name: default
data_files:
- split: train
path: data/train-*