For this LoRA I used 600 images of my favorite pixiv artists. The style is well suited for not shiny skin and thick thighs.
Trigger word: dbr34
V1: Basic version, may not work well with some poses.
V2: Different dataset. As a result there can be some less saturated images.
V3: Increased the dataset, fixed chromatic aberration.
V4: Removed bad images from the dataset, improved quality.
Description
Different dataset. As a result there can be some less saturated images. So expect V3.