metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: answer
dtype: string
- name: audio
dtype:
audio:
sampling_rate: 16000
splits:
- name: train
num_bytes: 345978043
num_examples: 2000
- name: validation
num_bytes: 82306957
num_examples: 500
- name: test
num_bytes: 542140907
num_examples: 3000
download_size: 975519022
dataset_size: 970425907
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*