metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: image
dtype: image
- name: category
dtype: string
splits:
- name: train
num_bytes: 3654746.156084502
num_examples: 100
download_size: 2239217
dataset_size: 3654746.156084502
configs:
- config_name: default
data_files:
- split: train
path: data/train-*