metadata
dataset_info:
features:
- name: video_id
dtype: string
- name: video
dtype: video
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: question
dtype: string
- name: candidates
list: string
- name: answer
dtype: string
splits:
- name: test
num_bytes: 19269949500
num_examples: 509
download_size: 19270618775
dataset_size: 19269949500
configs:
- config_name: default
data_files:
- split: test
path: data/test-*