metadata
dataset_info:
features:
- name: Prompt
dtype: string
- name: Chosen
dtype: image
- name: Rejected
dtype: image
splits:
- name: train
num_bytes: 298375432
num_examples: 100
download_size: 76722148
dataset_size: 298375432
configs:
- config_name: default
data_files:
- split: train
path: data/train-*