Hello everyone, It’s been a while! To be honest, I’ve been kinda stuck on Bleach lately, but fear not — I’ll also keep working on all of your requests. Today’s models took quite some time because I wasn’t really satisfied with the results. I ended up rolling back my training settings to the good old Kohya recommendation (constant with warmup), went back to the default 32/16 dim/alpha, and figured out my sweet spot for the number of steps — somewhere around 800 (batch size 2!). So now I can finally present you with some decent results... I hope. 😄Next one is Giselle Gewelle. Before we start, there’s one thing I should mention: this time, I actually didn’t use any gender tag [in training] at all, and that’s because...<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/2b91a439-7f63-41c2-9e92-31df30c0d5be/original=true/2b91a439-7f63-41c2-9e92-31df30c0d5be.jpeg" />...well, biologically speaking, that is. But you probably shouldn’t tell her that the way Yumichika did. 😅🤣<hr />Prompt: look at "Trigger Words" on the right side 👉Changelog: look at "About this Version" on the right side 👉<h1></h1>My Settings (forge neo)*:*note, the example images only use high-res fix.Checkpoint: waiNSFWIllustrious_v170 for IL, waiANIMA_v10Base10 for animaSampler: Euler a for both IL and AnimaScheduler: <s>Automatic</s> Karras for IL, Normal for AnimaSamplings Steps: 30 for both IL and AnimaHires.fix: R-ERSGAN 4x+Anime6B, 20 Hires steps for both IL and Anima, Denoising Strength: 0.3-0.5 for IL and 0.2-0.4 for animaExtensions (nice to have for little more Details):Adetailer (highly recommended)

Files with SHA256: cfc6b24dfdbd70e1de8a832daccfe5f8fe278637729e056c5cab6985ae1afaf0