As this is a merge based off <a target="_blank" rel="ugc" href="https://civarchive.com/models/257749/pony-diffusion-v6-xl">PonDiffusion v6</a> I HIGHLY recommend you read the description of that model for information regarding prompting, tags, and baked in characters.<h3 id="tldr-generation-settings:-3j2m5p0e7">TL;DR Generation Settings:</h3><ul><li>Sampler - Euler a / Euler a SGMUniform</li><li>Sampling Steps - 25</li><li>I upscale to 4x using <a target="_blank" rel="ugc" href="https://civarchive.com/articles/4560/upscaling-images-using-multidiffusion">this process </a>and the upscaler below:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/FacehugmanIII/4x_foolhardy_Remacri/blob/main/4x_foolhardy_Remacri.pth">4x_foolhardy_Remacri</a></li></ul></li><li>Dimensions - 832x1216, 1216x832, or 1024x1024</li><li>CFG Scale - 7 or higher</li><li>One or more passes of <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">ADetailer</a><ul><li>Some combination of the following for faces and eyes<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/207406/adetailer-face-finder-furry-anime-realistic-ddetailer">Facesv1.2</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/293448?modelVersionId=331333">Anzhc's Face Seg Prototype</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/342514?modelVersionId=383427">Anzhc's BeautifEYEr</a></li></ul></li></ul></li></ul>I use the following positive and negative prompt. I DO NOT USE ANY SOURCE TAGS (source_anime, source_furry etc) :<ul><li>score_9, score_8_up, score_7_up, BREAK {prompt}</li><li>NO NEGATIVE PROMPT OR EMBEDDINGS REQUIRED<ul><li>simply prompt for what you don't want to see as it comes up</li></ul></li></ul>See sample images for examples.Another personal merge, meant to generate bright, cartoony, and squishy characters. Like most of my models, this is intended to produce female characters and leans towards NSFW.V2: Rebaked many underlying models and remerged with different weights. This should hopefully increase the flexibility of the model.The baked-in artist styles for PD should work, but as this is effectively a style merge, they will defeat the purpose of this model.

Files with SHA256: 5a3989d6eeb9913fa634144a1359022ecd44b7b1b2e2aed0d16054983b97d851