metadata
dataset_info:
features:
- name: audio_duration
dtype: float64
- name: audio_token
sequence:
sequence:
sequence:
sequence: int64
- name: text
dtype: string
splits:
- name: train
num_bytes: 87042057050
num_examples: 2420047
download_size: 13484958846
dataset_size: 87042057050
configs:
- config_name: default
data_files:
- split: train
path: data/train-*