Today I was trying to create loras with an image, it practically turned out great but it is too strong and I was too lazy to edit the images to get a more adaptable lora so I preferred to train a character with about 30 images (yes, there are not many for her) XD