metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: speakers
sequence: string
- name: timestamps_start
sequence: float64
- name: timestamps_end
sequence: float64
splits:
- name: train
num_bytes: 2108045118
num_examples: 800
download_size: 1898909798
dataset_size: 2108045118
configs:
- config_name: default
data_files:
- split: train
path: data/train-*