The images are originally from this [fine-tuned dreambooth model](https://huggingface.co/jefsnacker/azzy). And it's just for study purpose to create this dataset so it'll be handy to load these images for further experiment.