metadata
dataset_info:
features:
- name: image
dtype: image
- name: text
dtype: string
- name: messages
dtype: string
splits:
- name: train
num_bytes: 31378
num_examples: 5
download_size: 33465
dataset_size: 31378
configs:
- config_name: default
data_files:
- split: train
path: data/train-*