metadata
dataset_info:
features:
- name: id
dtype: string
- name: input_image
dtype: image
- name: prompt
dtype: string
- name: verdict
dtype: string
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 3153794
num_examples: 20
download_size: 2238316
dataset_size: 3153794
configs:
- config_name: default
data_files:
- split: train
path: data/train-*