metadata
dataset_info:
config_name: small
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: whisper_transcript
sequence: int64
- name: input_length
dtype: int64
- name: labels
sequence: int64
splits:
- name: train
num_bytes: 2906406187
num_examples: 20900
download_size: 2866131051
dataset_size: 2906406187
configs:
- config_name: small
data_files:
- split: train
path: small/train-*