metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: transcription
dtype: string
- name: language
dtype: string
- name: source_dataset
dtype: string
- name: sample_id
dtype: string
- name: whisper_transcript
dtype: string
- name: eval_preds
list: int64
splits:
- name: train
num_bytes: 11863219
num_examples: 32
download_size: 10507484
dataset_size: 11863219
configs:
- config_name: default
data_files:
- split: train
path: data/train-*