<h1 id="mr-xotox-nasse-waende-sdxl-vcq441pds">MR-XOTOX-NASSE-WAENDE-SDXL</h1>More in About this versionModel 1.1 and 3.1 are not really trained models, but are provided with a Lora that contains my own photos that I once took! You can find more information about what I photograph at <a target="_blank" rel="ugc" href="http://www.der-zerfleischer.de">www.der-zerfleischer.de</a>At the moment I am working with Blip, Clip, Blip2 and Python to write my own local programmes for e.g. image captioning.Kohya_ss does not run at all on my computer, neither does image captioning via ComfyUI, so I sat down to develop such a programme for myself.You can find some videos on YouTube, but they are in German!<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist start="0"></iframe></div>

MR-XOTOX-NASSE-WAENDE-SDXL

First SDXL checkpoint of mine that was explicitly set to DPM++ 3M SDE Karras.In addition to the first version, 2 additional realistic checkpoints were added. So the 2nd version theoretically consists of 5 checkpoints. I added another feature via UNET to achieve the sampler / scheduler variant at all!IMPORTANT!!!!!!!!Sampler: DPM++ 3M SDEScheduler: KarrasCFG: 2.5-7 (I have not tried 7+)👋 CLIP-SKIP: -3 ComfyUI / 3 A1111 👋Give me some feedback if the results are okay, and maybe a thumbs up ;-) 👍

Files with SHA256: 1a6b42d5dd3566bd82673d2914b84bdede9dd479be66959111f5820b34f41591