metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: image
- name: rejected
dtype: image
splits:
- name: train
num_bytes: 64513220
num_examples: 20
download_size: 64517866
dataset_size: 64513220
configs:
- config_name: default
data_files:
- split: train
path: data/train-*