metadata
dataset_info:
features:
- name: image_path
dtype: string
- name: caption
dtype: string
- name: question
dtype: string
- name: image
dtype: image
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 75527314
num_examples: 745
download_size: 74905983
dataset_size: 75527314
configs:
- config_name: default
data_files:
- split: train
path: data/train-*