metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcription
dtype: string
- name: w2v2_transcription
dtype: string
- name: WER
dtype: int64
- name: sum
dtype: float64
splits:
- name: train
num_bytes: 16003127322.525991
num_examples: 194747
- name: test
num_bytes: 1778161779.8073397
num_examples: 21639
download_size: 17552533348
dataset_size: 17781289102.333332