metadata
dataset_info:
features:
- name: file
dtype: string
- name: is_valid
dtype: bool
- name: language
dtype:
class_label:
names:
'0': de
- name: speaker_id
dtype: string
- name: gender
dtype:
class_label:
names:
'0': MALE
'1': FEMALE
'2': OTHER
'3': NAN
- name: keyword
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: pitch
dtype: float64
splits:
- name: train
num_bytes: 9571402808.75
num_examples: 295802
- name: validation
num_bytes: 892350026.75
num_examples: 27578
- name: test
num_bytes: 891750997.125
num_examples: 27559
download_size: 3176016351
dataset_size: 11355503832.625
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*