metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: corpus
dtype: string
splits:
- name: train
num_bytes: 33669688.235242225
num_examples: 16324
download_size: 15127813
dataset_size: 33669688.235242225
configs:
- config_name: default
data_files:
- split: train
path: data/train-*