metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: task_id
dtype: int8
- name: task
dtype: string
- name: source_dataset
dtype: string
- name: source_view
dtype: string
- name: prompt
dtype: string
- name: response
dtype: string
- name: image
list: image
- name: coord_type
dtype: string
- name: n_boxes
dtype: int32
- name: n_points
dtype: int32
- name: official_parsable
dtype: bool
splits:
- name: train
num_bytes: 77265580291
num_examples: 2007237
download_size: 677031637090
dataset_size: 77265580291