metadata
dataset_info:
features:
- name: file
dtype: string
- name: audio
dtype: audio
- name: label
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 11845511188
num_examples: 27002
- name: validation
num_bytes: 1304884322
num_examples: 3001
download_size: 12631265875
dataset_size: 13150395510
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*