Sign In

Ryoko Kui Portraits (Dungeon Meshi etc.) [Pony XL]

23
158
1.4k
13
Updated: Jun 29, 2024
style
Verified:
SafeTensor
Type
LoRA
Stats
158
1,428
Reviews
Published
Jun 29, 2024
Base Model
Pony
Usage Tips
Clip Skip: 2
Strength: 0.9
Training Images
Download
Hash
AutoV2
22128140D9
default creator card background decoration
AnyKey's Avatar
AnyKey

Makes portraits similar in style to Ryoko Kui's Baldurs Gate shots.

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.