Hana

This LoRA was created to generate a waifu that resembled a beautiful girl I saw on a trip to Ireland. The hairstyle and clothes were similar to how I tried to portray her with this model. The LoRA tends to produce images from a top-down perspective. It was trained starting from a single image, from which I created videos using Kling, and then used those to generate the training images. Unfortunately, there wasn’t much variety in the framing angles, so the tendency is to generate a top-down POV view.