<span style="background-color:rgb(23, 23, 23);color:rgb(183, 183, 183);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">this model is good at creating that ideal Brunette that you dream of, it is really good at body formation and face also but has trouble when creating a pose where she has to look back at the camera, with her backside showing. not all are failures, but works better for certain aspects than others depending on the scheduler. haven't found that absolute sweet spot yet, but Beta and dpm++_sde seems to work best with 25-30 steps and 3.5-6.0 cfg. the lora is trained on the Lustify v7 checkpoint and was trained with a completely fictional data set from created from one of my other Loras <a target="_blank" rel="ugc" href="https://civarchive.com/models/2357911/alex-vibes-v3">Alex vibes V3 - V1 | Stable Diffusion XL LoRA | Civitai </a> creates even better pictures with a beach setting because of similar images for its training

Anya DayZz SDXL

use Dpm++_sde and Beta for beast results with a 0.65 lora strength.

Files with SHA256: e6664eeeac6df7578a281c59c56207c43a524cc6e0507a6fd0a1351e33633846