metadata
dataset_info:
features:
- name: text_input
dtype: string
- name: audio_input
sequence: float64
- name: audio_input_sr
dtype: int64
- name: text_output
dtype: string
- name: audio_output
sequence: float16
- name: audio_output_sr
dtype: int64
splits:
- name: train
num_bytes: 16651789898
num_examples: 7605
download_size: 12667501323
dataset_size: 16651789898
configs:
- config_name: default
data_files:
- split: train
path: data/train-*