metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: text
dtype: string
- name: transliterated
dtype: string
- name: latents
list:
list: float32
- name: audio
dtype:
audio:
sampling_rate: 16000
decode: false
- name: duration
dtype: float64
splits:
- name: train
num_bytes: 63230137152
num_examples: 746335
download_size: 63184967919
dataset_size: 63230137152