<h2 id="refined-2d-8vne246ku">[Refined 2D]</h2>This is a merged model specifically designed to have a crisp, highly detailed and dynamic digital art-style that works with whatever subject you're trying to prompt.Update: I've re-merged this model and baked in the correct vae! Images should no longer be washed out without one.Follow the base-prompt provided below for the best results!(I personally use the Forge Webui by Illyasviel)Recommended Settings:<ul><li>Sampler: Euler A</li><li>CFG Scale: 4.5</li><li>Steps: 40-60</li><li>Clip Skip: 2</li><li>Resolution: Standard SDXL resolutions work well, but try going above 1024x1024 for better results! This model can give more detailed results with a resolution of 1348x1524 for instance.</li></ul>Try enabling:<ul><li>FreeU</li><li>SelfAttentioinGuidance</li><li>Kohya HRFix (downscale factor 1.5, end percent 0.1)</li></ul>These settings generally make the image adhere more to the prompt, be sure to use the Forge Webui by Illyasviel to get the results provided in the example images!Recommended Prompt:very awa, masterpiece, best quality, newest, highres, absurdres, 3d, huffslove,BREAK[describe your character, pose and actions here]BREAKdynamic angle, dynamic pose, motion blur, depth of field, detailed background, environment, [describe the background here]Recommended Negative:worst aesthetic, bad quality, old, watermark, text, monochrome, solo, bad anatomyPlease share any creations you might make with this model! And try out the provided prompt for some great dynamic results, it's perfect for using with Wildcards to generate a great assortment of images without having to change out the base prompt all the time!

Refined 2D

This is the first and most likely final version of this model. It's highly dynamic and I find the style close to perfect in most cases. I will attempt to make another merged model in this style once the next big diffusion model releases.

Files with SHA256: 2b04414f809fddc0732c0d694f1d08e964e99395dca42aeb07511adc112de367