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: 52578533819.2
num_examples: 45876
- name: test
num_bytes: 6572889777.947336
num_examples: 5735
- name: valid
num_bytes: 6571743676.852664
num_examples: 5734
download_size: 21994326235
dataset_size: 65723167273.99999
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*