Use 0.7 - 1.0 strength for best results.
Trigger word: oseronstyle
Supporting words: 1girl, solo, transformation, anthro, feral, sweaty, fat, chubby
The LoRA was trained primarily on pigs, frogs, tanukis, and mice, but functions well on all kinds of animals such as cows, dogs, monkeys, donkeys, and more. It does very well at transformation sequences.
Many of the trained images were sketches, so you may wish to put "monochrome" or "greyscale" in negative. In general, issues should not arise from this, though.
Showcase images have not been edited or inpainted in order to demonstrate raw output. Some of the images were generated with the assistance of this LoRA which for whatever reason didn't auto-detect as being used, so it doesn't show up in resources. It isn't necessary to use it, just a LoRA whose visual effect I enjoy.
For the sake of matters of copyright and consent, I do not train directly on artists' works. Every image trained on is either a lookalike drawing by another artist or an image generated by another model/LoRA.
Description
FAQ
Comments (5)
This is pretty fab!
I cant get this to output sequences like in your preview images.
My preview images were generated offsite on my own installation, so it's possible they won't quite match what can be done on Civitai. However, I just went to the first dog image and chose to remix it, and it managed to give me something similar.
You should be able to click Remix on this image, make sure "mature" generation is turned on, scroll to the bottom, choose "random" instead of "custom," then click Generate, and you will get something similar to it: https://civitai.com/images/23577193
If you're doing it on your own off of Civitai, make sure you're using PonyV6, this LoRA at 0.8 strength, with a prompt identical or similar to the one I used. Make sure you use the words "sequence, transformation sequence," I don't know if both are required but it seems to help.
If you don't have trouble making a nearly-identical image when using the same prompt, but do have trouble when trying to write your own prompt from scratch, try starting with mine and then change just one small thing, then another, then another, and find out which thing you removed seems to make it stop doing a sequence.
Hi, I am also using my own local setup, I dont generate on the website and no matter what ive tried I cant get the images to go in a sequence like this
@LocalYokel Not sure what the issue is then. I guess just try to make every parameter as close as possible to the examples, including resolution. I think if you generated in too low of a resolution like 512x768 you might experience that problem. If you're using other LoRAs or embeddings, try taking them out, especially if you have a big negative prompt with lots of stuff in it.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.