metadata
dataset_info:
features:
- name: input_ids
sequence: int64
- name: length
dtype: int64
splits:
- name: train
num_bytes: 3960296272
num_examples: 988730
- name: valid
num_bytes: 216896820
num_examples: 54753
download_size: 951004341
dataset_size: 4177193092
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*