metadata
dataset_info:
features:
- name: question_id
dtype: string
- name: video_path
dtype: string
- name: area
dtype: string
- name: conversations
list:
- name: role
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 3009379
num_examples: 7392
- name: test
num_bytes: 7765932
num_examples: 19140
download_size: 792760
dataset_size: 10775311
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*