My LoRA are made using difference training and a dataset of two images. I'm also too lazy to re-download base noob to train off of so I'm sorry if there are issues.