metadata
dataset_info:
features:
- name: duration
dtype: float64
- name: audio
dtype: audio
- name: transcription
dtype: string
- name: transcript_length
dtype: int64
- name: predictions
dtype: string
- name: wer
dtype: float64
- name: cer
dtype: float64
splits:
- name: test
num_bytes: 221335492.2749684
num_examples: 393
download_size: 271364504
dataset_size: 221335492.2749684
configs:
- config_name: default
data_files:
- split: test
path: data/test-*