metadata
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: text
dtype: string
- name: actions
sequence:
sequence: float32
- name: path
dtype: string
splits:
- name: train
num_bytes: 110654468
num_examples: 769
download_size: 115337909
dataset_size: 110654468
configs:
- config_name: default
data_files:
- split: train
path: data/train-*