metadata
dataset_info:
features:
- name: audio
struct:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: original_audio
struct:
- name: bytes
dtype: binary
- name: path
dtype: string
- name: sentence
dtype: string
splits:
- name: train
num_bytes: 16552493607
num_examples: 7232
- name: validation
num_bytes: 2069061700
num_examples: 904
- name: test
num_bytes: 2069061700
num_examples: 904
download_size: 19905886970
dataset_size: 20690617007
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*