metadata
dataset_info:
features:
- name: context
dtype:
audio:
sampling_rate: 192000
- name: instruction
dtype: string
- name: speech_instruction
dtype: string
- name: answer
dtype: string
splits:
- name: test
num_bytes: 228634291
num_examples: 100
download_size: 197814930
dataset_size: 228634291
configs:
- config_name: default
data_files:
- split: test
path: data/test-*