metadata
dataset_info:
features:
- name: context
dtype:
audio:
sampling_rate: 16000
- name: instruction
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 3425832242
num_examples: 20000
- name: test
num_bytes: 373560758
num_examples: 2000
download_size: 3715675082
dataset_size: 3799393000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*