metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: ratio_char_token
dtype: float64
splits:
- name: train
num_bytes: 57725193148
num_examples: 5300000
download_size: 24092790473
dataset_size: 57725193148
configs:
- config_name: default
data_files:
- split: train
path: data/train-*