metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcription
dtype: string
- name: predictions
dtype: string
- name: gender
dtype:
class_label:
names:
'0': male
'1': female
'2': other
- name: wer
dtype: float64
- name: cer
dtype: float64
splits:
- name: test
num_bytes: 86837576.05660377
num_examples: 80
download_size: 81374742
dataset_size: 86837576.05660377
configs:
- config_name: default
data_files:
- split: test
path: data/test-*