metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: speaker
dtype: string
- name: text
dtype: string
- name: target
dtype: string
- name: duration
dtype: float64
- name: augmentation
struct:
- name: clean
dtype: bool
- name: noise_type
dtype: string
- name: snr_db
dtype: float64
- name: reverb
dtype: bool
- name: speed_factor
dtype: float64
- name: lead
dtype: float64
- name: trail
dtype: float64
splits:
- name: train
num_bytes: 2679017305
num_examples: 10228
download_size: 3325787460
dataset_size: 2679017305
configs:
- config_name: default
data_files:
- split: train
path: data/train-*