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: 53657396733.81804
num_examples: 47042
- name: test
num_bytes: 6708030062.318436
num_examples: 5881
- name: valid
num_bytes: 6706889434.863527
num_examples: 5880
download_size: 22095060839
dataset_size: 67072316231
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*