metadata
dataset_info:
features:
- name: scene_id
dtype: string
- name: situation
dtype: string
- name: question
dtype: string
- name: ground_truth_answer
dtype: string
- name: ground_truth_reasoning
dtype: string
- name: time_span
list: int32
- name: image_1
dtype: image
- name: image_2
dtype: image
- name: image_3
dtype: image
- name: image_4
dtype: image
- name: image_5
dtype: image
splits:
- name: train
num_bytes: 7370351
num_examples: 1
download_size: 7378306
dataset_size: 7370351
configs:
- config_name: default
data_files:
- split: train
path: data/train-*