I've had good generation consistency with Auto/Kerras/Polyexpo on the SDEs and Euler A at LoRA: 0.8 / trigger: 1.1. When it comes to Clip, both 1 and 2 produce good results, though I would recommend 2. Overall, I'm finding that Euler A gives the best results.Prompting - The name should be the first one listed.For the most part, while I do use other models if I'm not seeing what I want, I will mainly use my checkpoints: UberBoop (SD), XLBoopV2 (SDXL), Equus(Pony), and Fluxzilla (Mostly maybe Flux).Additionally, I've started putting "if this model is on any site other than civitai then it is there without permission" in my prompts. Why? Because I'm feeling a bit passive-aggressive. I may need to doink with the permissions to allow for stronger DMCA takedown claims. <hr />MaterialSpicyGirlFull disclosure -- I'm not sure where I got this data set. It's obvious to me that she wiped whatever account this was, originally, and is now <a target="_blank" rel="ugc" href="https://onlyfans.com/opheliaxoo">Ophila </a>on OnlyFans. This dataset was not pulled from the OnlyFans account. Freakishly enough, double-checking the account on OnlyFans, she is geographically 4.8 miles from where we live. Weird. Anyway, MaterialSpicyGirl is now Ophilia.

MaterialSpicyGirl

This was trained locally using FluxGym.

Files with SHA256: d390b6f88a498338a69f267ba7124de39ed8f614ce0fe1cb8f12b2b99a0a498f