metadata
dataset_info:
features:
- name: video_id
dtype: string
- name: scene_id
dtype: string
- name: qa
dtype: string
- name: timestamps
dtype: string
- name: timecoded_cc
dtype: string
- name: video_url
dtype: string
splits:
- name: train
num_bytes: 289368960
num_examples: 66410
download_size: 86484129
dataset_size: 289368960
configs:
- config_name: default
data_files:
- split: train
path: data/train-*