metadata
dataset_info:
features:
- name: image
dtype: image
- name: class_name
dtype: string
- name: prompt
dtype: string
- name: expected_response
dtype: string
splits:
- name: train
num_bytes: 840934053.759
num_examples: 4803
download_size: 837543226
dataset_size: 840934053.759
configs:
- config_name: default
data_files:
- split: train
path: data/train-*