metadata
dataset_info:
features:
- name: youtube_id
dtype: string
- name: audiocaps_id
sequence: string
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: caption
sequence: string
- name: label
sequence: string
- name: hallucination_label
sequence: string
splits:
- name: test
num_bytes: 607000900
num_examples: 957
download_size: 594451576
dataset_size: 607000900
configs:
- config_name: default
data_files:
- split: test
path: data/test-*