metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: path
dtype: string
- name: condition_on_prev
dtype: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 1064819610.125
num_examples: 1215
- name: validation
num_bytes: 267679544
num_examples: 306
download_size: 1329977753
dataset_size: 1332499154.125
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*