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: 49828978532.05408
num_examples: 43076
- name: test
num_bytes: 6229200700.9729595
num_examples: 5385
- name: valid
num_bytes: 6229200700.9729595
num_examples: 5385
download_size: 21817490918
dataset_size: 62287379934
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*