metadata
dataset_info:
features:
- name: scene_id
dtype: string
- name: frame_id
dtype: string
- name: qa_id
dtype: string
- name: scene_description
dtype: string
- name: category
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: context
dtype: string
- name: tag
list: int32
- name: cluster
dtype: string
- name: layer
dtype: string
- name: con_up
dtype: string
- name: con_down
dtype: string
- name: image
dtype: image
- name: all_camera_images
struct:
- name: cam_front
dtype: string
- name: cam_front_left
dtype: string
- name: cam_front_right
dtype: string
- name: cam_back
dtype: string
- name: cam_back_left
dtype: string
- name: cam_back_right
dtype: string
- name: key_objects
dtype: string
- name: num_key_objects
dtype: int32
splits:
- name: train
num_bytes: 777469
num_examples: 10
download_size: 93335
dataset_size: 777469
configs:
- config_name: default
data_files:
- split: train
path: data/train-*