metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: Transcription
dtype: string
- name: TranscriptionTranslated
dtype: string
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 3528025047.107
num_examples: 4259
- name: test
num_bytes: 385823300
num_examples: 464
download_size: 3893685259
dataset_size: 3913848347.107
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*