metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: logits
sequence: float32
splits:
- name: train
num_bytes: 511855989.7994287
num_examples: 15966
download_size: 435346692
dataset_size: 511855989.7994287
configs:
- config_name: default
data_files:
- split: train
path: data/train-*