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: 3247616054
num_examples: 2016
- name: validation
num_bytes: 723406204
num_examples: 408
- name: test
num_bytes: 735869160
num_examples: 408
download_size: 3506142041
dataset_size: 4706891418
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*