metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: labels
dtype: string
- name: input_features
sequence:
sequence: float32
splits:
- name: train
num_bytes: 50374532662.7424
num_examples: 43631
- name: test
num_bytes: 6296960902.628798
num_examples: 5454
- name: valid
num_bytes: 6296960902.628798
num_examples: 5454
download_size: 21881898008
dataset_size: 62968454468
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*