Manuscript drolleries

详情

下载文件 (1)

关于此版本

This need some specific instruction in the prompt to work well. Embeddings will interfere with your prompt. Note: this is a LoRA based on marginalia, AKA doodles on the side of the text in illuminated medieval manuscripts, 'text' in the negatives could be omitted if you want and in the positives could even trigger the LoRA.

Tags that work: fine art parody, colored ink \(medium\), branch, leaves, text, stylized anatomy, paper texture \(traditional media\), faded background, simple background, antiqued

And maybe put things like 3d, render, volumetric shadowing, photorealistic in the negatives.

If you want to use this in some IL based checkpoint, start more or less from strenght 2

模型描述

SDXL, this time with prodigy

A lot of images of grotesques, drolleries, marginalia and doodles from illuminated manuscripts went inside, there were some magic and poof: you have this LoRA.

See the notes for how to use it, as it alone isn't enough: you need a stylized 2d style as a base, or else you will not really obtain medieval doodles. I tried the style LoRA of tehalex86 and I found the PEN REALISM version works wonders (read its instructions too).

此模型生成的图像

排序

模型类型	LORA
基础模型	SDXL 1.0
发布时间	2025-08-05