metadata
dataset_info:
features:
- name: unique_id
dtype: string
- name: object_id
dtype: string
- name: category
dtype: string
- name: task_type
dtype: string
- name: question
dtype: string
- name: cot
dtype: string
- name: answer
dtype: string
- name: images
list: image
- name: point_cloud
dtype:
array2_d:
shape:
- 8192
- 3
dtype: float32
splits:
- name: train
num_bytes: 13835735904.175999
num_examples: 68613
- name: validation
num_bytes: 1772591333.415
num_examples: 8835
- name: test
num_bytes: 1763319710.704
num_examples: 8832
download_size: 9255330360
dataset_size: 17371646948.295
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*