metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: sentence
dtype: string
- name: path
dtype: string
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 17610490756.96
num_examples: 95808
- name: test
num_bytes: 4385076703.671468
num_examples: 23951
download_size: 21640531900
dataset_size: 21995567460.631466
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*