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: 49598093483.581436
num_examples: 42896
- name: test
num_bytes: 6200917925.970888
num_examples: 5363
- name: valid
num_bytes: 6199761685.4476795
num_examples: 5362
download_size: 21776495778
dataset_size: 61998773095
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*