metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: instruction
dtype: string
- name: label
dtype: string
- name: taskname
dtype: string
splits:
- name: train
num_bytes: 223444431
num_examples: 1000
download_size: 192388489
dataset_size: 223444431
configs:
- config_name: default
data_files:
- split: train
path: data/train-*