metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: k
dtype: int64
- name: components
list:
- name: first_word_start_sec
dtype: float64
- name: speaker_id
dtype: string
- name: utterance_id
dtype: string
- name: utterance_length_sec
dtype: float64
splits:
- name: train
num_bytes: 6405930308
num_examples: 20000
- name: test
num_bytes: 640592871
num_examples: 2000
download_size: 7012732331
dataset_size: 7046523179
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*