metadata
license: mit
dataset_info:
features:
- name: image
dtype: image
- name: image_url
dtype: string
- name: qwen_caption
dtype: string
- name: verified_caption
dtype: string
- name: model_caption
dtype: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 349884039
num_examples: 164
download_size: 349105732
dataset_size: 349884039
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
language:
- en
- lt
Captions for the images are first generated with Qwen3-VL.
This is a project for a university project. The project will not be used outside of educational purposes.
Reach out if you have any complaints.