Adam's Artwork Style

A LoRA trained on images in my preferred artwork style. It should work with most 1.5 models, but I've found that it works best with the Dungeons N Waifu's model with a LoRA weight between 0.25 and 0.5, and a CFG Scale of 4.5

Smaller faces and hands generated in this style tend to be a little scrungy and might require inpainting, since the GFPGAN style conflicts with the painterly style (see the dancing bohemian woman for example)

Upscaling by x1.5 using Lanczos for 10 steps with a denoising strength of 0.3 usually fixes up the image nicely, though YMMV

All example pictures are done in Dungeons N Waifu's without negative prompting except where necessary

Some examples of prompts to try:

• AJAWS, character concept art of a beautiful person downloading the most beautiful lora <lora:15_AJAWS:0.5> Steps: 30, Sampler: DDIM, CFG scale: 4.5, Seed: 145, Size: 512x768, Model hash: cb79fb0155

• AJAWS, a man at an art gallery crying tears of joy at art made by the most beautiful lora <lora:15_AJAWS:0.5>
  Steps: 30, Sampler: DDIM, CFG scale: 4.5, Seed: 145, Size: 768x512, Model hash: cb79fb0155