metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: rejected
dtype: image
- name: chosen
dtype: image
- name: metadata
dtype: string
splits:
- name: train
num_bytes: 2035517350
num_examples: 651
download_size: 2035514393
dataset_size: 2035517350
configs:
- config_name: default
data_files:
- split: train
path: data/train-*