metadata
dataset_info:
features:
- name: image
dtype: image
- name: caption
dtype: string
- name: image_name
dtype: string
- name: source_folder
dtype: string
- name: responses
dtype: string
splits:
- name: train
num_bytes: 36215686
num_examples: 1000
download_size: 34576094
dataset_size: 36215686
configs:
- config_name: default
data_files:
- split: train
path: data/train-*