metadata
dataset_info:
features:
- name: file
dtype: string
- name: audio
dtype: audio
- name: instruction
dtype: string
- name: label
dtype: string
splits:
- name: train
num_bytes: 4844413128.351
num_examples: 16623
- name: validation
num_bytes: 679648410.816
num_examples: 2392
download_size: 5172632764
dataset_size: 5524061539.167
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*