metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: image
dtype: image
- name: rejected
dtype: string
- name: chosen
dtype: string
splits:
- name: train
num_bytes: 1390793478
num_examples: 10000
download_size: 1387930416
dataset_size: 1390793478
configs:
- config_name: default
data_files:
- split: train
path: data/train-*