metadata
dataset_info:
features:
- name: machine_unit
sequence:
sequence: int64
- name: x-vector
sequence:
sequence:
sequence: float32
- name: text
sequence: string
- name: text_with_pad
sequence: string
- name: user_audio_path
dtype: string
- name: not_aligned_percentage
dtype: float64
- name: duration
dtype: int64
splits:
- name: train
num_bytes: 93319513179
num_examples: 760438
download_size: 13637561128
dataset_size: 93319513179
configs:
- config_name: default
data_files:
- split: train
path: data/train-*