metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: phonemes
sequence: string
splits:
- name: train
num_bytes: 61041277210
num_examples: 5159931
download_size: 21328828261
dataset_size: 61041277210
configs:
- config_name: default
data_files:
- split: train
path: data/train-*