metadata
dataset_info:
features:
- name: question
dtype: string
- name: question_type
dtype: string
- name: image_src
dtype: string
- name: question_topic
dtype: string
- name: gt_answer
dtype: string
- name: model_answer
dtype: string
- name: image_content
sequence: string
- name: image_id
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 175937176
num_examples: 96
download_size: 175127368
dataset_size: 175937176
configs:
- config_name: default
data_files:
- split: train
path: data/train-*