metadata
dataset_info:
features:
- name: images
sequence: image
- name: caption
dtype: string
- name: cdl
dtype: string
- name: proID
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 44135754.59
num_examples: 4862
- name: validation
num_bytes: 13606327.6
num_examples: 1400
download_size: 58411294
dataset_size: 57742082.190000005
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*