metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: images1
list: image
- name: images2
list: image
- name: args1
dtype: string
- name: args2
dtype: string
splits:
- name: train
num_bytes: 141973194
num_examples: 21
download_size: 141945226
dataset_size: 141973194
configs:
- config_name: default
data_files:
- split: train
path: data/train-*