Makeup Loras

Introduction

Are you looking for more precise control over eyeliner styles in your AI art? The Eyeliner v1 LoRA, part of the Makeup Loras series, is designed to give you just that. It's particularly useful for users of the ZImageTurbo base model who want to avoid the default "cat eye makeup" that often appears with heavy eyeliner prompts. This LoRA helps you achieve a wider range of eyeliner looks, from subtle to bold, with greater consistency.

How to Use

Getting started with the Eyeliner v1 LoRA is simple, as it doesn't require any specific trigger word. Just add it to your prompt and let it work its magic.
For most eyeliner effects, we recommend setting the LoRA's weight (strength) between 0.5 and 1.0. Starting at 0.7 and adjusting up or down is a good approach to find the perfect intensity for your desired look.
A unique feature of this LoRA is its ability to influence makeup application in reverse. By using negative weights, specifically from -0.5 to -1.0, you can prompt for a natural or "nude" face, effectively removing any makeup from your generated images.
This LoRA was specifically trained to address a common behavior of the ZImageTurbo checkpoint: its tendency to generate "cat eye makeup" when heavy eyeliner is requested. While not 100% effective, Eyeliner v1 significantly helps in reducing or negating this specific style about half the time, giving you more variety and control over your character's eye makeup.