This checkpoint is the successor of Scribbler XL. It is more versatile and more colorful. It is a mix between a merge and trained checkpoint. I used my <a target="_blank" rel="ugc" href="https://civarchive.com/models/492548">Scribbler XL</a> checkpoint, added &lt;5% from <a target="_blank" rel="ugc" href="https://civarchive.com/models/140737">AlbedoBase XL v2.1</a> (a very versatile checkpoint) and finalized it with thousands of images i created with a LoRA to get the style. I really recommend looking into the example pictures to get a start how this thing works. It is not made for photorealistic stuff, despite it can do real environments but it's not the strength of this model. Add 'photo', 'photorealistic', 'photoshot' or similar words to the negative prompt. Oh, and it has a flaw of course. Sometimes i got triple legs, so i checked the whole dataset and removed manually everything which might cause this effect (even some shadows!). So beware of the tripods... rarely... See it as a personal watermark ;) Edit: This Version has now a VAE baked in. Thanks to Khonar.

Files with SHA256: e1a61d295f3063ee005f2dbb6ac1b9d5dbd403fc4f1022f7cc904cc16c2139c1