metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: word
dtype: string
- name: phoneme
dtype: string
- name: file
dtype: string
- name: instruction
dtype: string
- name: label
dtype: int64
splits:
- name: test
num_bytes: 104506606.652
num_examples: 6598
download_size: 103899602
dataset_size: 104506606.652
configs:
- config_name: default
data_files:
- split: test
path: data/test-*