metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: logits
sequence: float32
splits:
- name: train
num_bytes: 661668280.9400865
num_examples: 20639
download_size: 602058081
dataset_size: 661668280.9400865
configs:
- config_name: default
data_files:
- split: train
path: data/train-*