metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: speaker
dtype: string
- name: variety
dtype:
class_label:
names:
'0': Arabic
'1': Hindi
'2': Korean
'3': Mandarin
'4': Spanish
'5': Vietnamese
- name: text
dtype: string
- name: blanked
dtype: string
- name: options
sequence: string
- name: actual
dtype: string
splits:
- name: train
num_bytes: 115715892
num_examples: 450
download_size: 115683759
dataset_size: 115715892
configs:
- config_name: default
data_files:
- split: train
path: data/train-*