Makes portraits similar in style to Ryoko Kui's Baldurs Gate <a rel="ugc" href="https://forum.neverwintervault.org/t/baldurs-gate-protraits-by-ryoko-kui/2635">shots</a>.How to use Prompt "1boy" or "1girl" followed by the "portrait" tag and then whatever else you want to include. Due to the dataset this LoRA works great for lower resolution portraits suitable for character sheets, headshots for in game dialogue and such. Consider content outside this like full/high resolution images, full body shots, scenery and so on completely unsupported.Recommended workflow is simply generating at 768*1024 and then scale the image down to the whatever resolution works best for your use case/game.There's a wide variety in the images trained on so most content should work well. A special note is that there's a fairly even selection of 1boy/1girl providing good support to prompt male characters as well - young, old, manly, facial haired and so on.For more specific prompt tips what content tags you can use please see the zip with training tags.MISC Trained on Pony SDXL at 1024 resolution with a dataset of 37 images. All sample images are raw txt2img upscales, no inpainting or helper embeddings used, e.g - your outputs should look better with more than minimum effort. The training was done on a small subset of the artists images, as such the output of this LoRA is not an accurate representation neither in quality or style of the source material.

Files with SHA256: 22128140d94d8d403e029e0b6e7c885a7b7fab68df7a792d3fb1889f30e76b04