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: 50910604189.62563
num_examples: 44170
- name: test
num_bytes: 6364689978.155145
num_examples: 5522
- name: valid
num_bytes: 6363537372.219224
num_examples: 5521
download_size: 21941784881
dataset_size: 63638831540
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
- split: valid
path: data/valid-*