Topsy ʎʌɹnꓕ
Details
Download Files
Model description
This LoRA was trained on a variety of images that were turned upside-down without telling Stable Diffusion they were upside-down. Because most if not all checkpoints are trained on normally oriented images, don't expect most results to be realistically inverted but instead a disorienting and often mutated mixture.
You might not spot just how weird your generations with this LoRA are until you flip them "right side up."



