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: 48795616276.267365
num_examples: 42046
- name: test
num_bytes: 6099742166.866319
num_examples: 5256
- name: valid
num_bytes: 6099742166.866319
num_examples: 5256
download_size: 21651808856
dataset_size: 60995100610
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*