modelres + beta schedulernoob eps finetunes -&gt; vpred + noob vpred finetunes -&gt; various incest maneuvers (illPersonalMerge_v30, noobieater_v30, obsessionIllustrious_v3, obsessionIllustrious_vPredV10, catTowerNoobaiXL_v15Vpred, noobaiCyberfixV2_10vpredPerp, EasyFluffXLVpred, some loras and manual layer scaling)<hr />7/9/2025 READ THE VERSION DESCRIPTIONS ON THE RIGHTversions based on incest from february until june1.5 - 1.0 + various finetunes, illust 1.0, illust 1.1, animagine 4.0, unreleased noob vpred1.69 - same + routwei + nhl001 lora - &gt; STYLISTIC DOGSHIT2048a - 1.5 with more of illust 1.1 and 2.0 -&gt; THIS IS AN INGREDIENT ONLY MODEL2048b - 1.5 with less of illust 1.1 and 2.0 -&gt; SERVICABLE 2048x GENSalternate 2048b equivalents here, they're better or worse sometimes<a target="_blank" rel="ugc" href="https://huggingface.co/Enferlain/juice/tree/main/noob/2048">https://huggingface.co/Enferlain/juice/tree/main/noob/2048</a>someone please fix model staregeand join sd mecha discord

ΣΙΗ

more stable at 2048x1536/1536x2048 but at the cost of basically everything else. can barely be called vpred, but enough to maybe train or do some more incest on it. I got tired of merging layers 1 by 1 and testing so my last few attempts until illust 3.5 will be training related

Files with SHA256: 7fe8cbef6aa11090b8499d29d08e1bc167abf27d1a3b35ee20605e7a37907586