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: 53038769822.08567
num_examples: 46399
- name: test
num_bytes: 6629989115.457163
num_examples: 5800
- name: valid
num_bytes: 6629989115.457163
num_examples: 5800
download_size: 22209915957
dataset_size: 66298748052.99999
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*