metadata
dataset_info:
features:
- name: merged_phonemes
dtype: string
- name: input_ids
sequence: int32
- name: phonemes
sequence: string
splits:
- name: train
num_bytes: 24927984204
num_examples: 1949339
download_size: 11978919561
dataset_size: 24927984204
configs:
- config_name: default
data_files:
- split: train
path: data/train-*