metadata
dataset_info:
features:
- name: context
dtype: audio
- name: instruction
dtype: string
- name: speech_instruction
dtype: string
- name: answer
dtype: string
- name: sound_tokens
dtype: string
splits:
- name: test
num_bytes: 2996060
num_examples: 20
download_size: 2643138
dataset_size: 2996060
configs:
- config_name: default
data_files:
- split: test
path: data/test-*