metadata
dataset_info:
features:
- name: source_id
dtype: int64
- name: split
dtype: string
- name: image_path
dtype: string
- name: output_image_path
dtype: string
- name: source
dtype: string
- name: scene
dtype: string
- name: class
dtype: string
- name: question
dtype: string
- name: options
list:
list: int64
- name: answer
dtype: int64
- name: image
dtype: image
- name: output_image
dtype: image
splits:
- name: train
num_bytes: 1840404508.856
num_examples: 2392
download_size: 1846922069
dataset_size: 1840404508.856
configs:
- config_name: default
data_files:
- split: train
path: data/train-*