(Request)
Character + Style
Tested and works fine between (0.8 & 1.2) should work with any pony model. But i do recommend anime style pony models !!
Main prompts:
revy, purple hair, brown eyes, ponytail, shoulder tattoo,
revy, purple hair, brown eyes, long hair, shoulder tattoo,
tank top, midriff, holster, short shorts, belt, fingerless gloves,
Description
FAQ
Comments (12)
WOMP WOMP :D ohhh myyy
pretty good lora, but why wipe the metadata on the previews?
Do you mean like the prompts, checkpoint, etc?
If so, i generate these images outside of CivitAI and i use inpainting. After that, i do a few touch ups with photoshop, because of that it wipes the metadata :(
If you have any other question feel free to ask :) !!
@Taz_ia oh I see. That's fine. I'd like to know the prompt for the image you generated (any one), and the checkpoint/style lora (if any).
@adshfbiulaerfb For the checkpoint, i am using a mix i made of multiple model and style loras, so i dont really remember all the ressources i used.
style lora i recommend the following:
https://civitai.com/models/264290?modelVersionId=333590
https://civitai.com/models/396772/modern-anime-screencap-style-pony-xl
https://civitai.com/models/497276/nsfw-anime-screenshot-zuharu-or-style-lora-or-pony-xl
or any other anime style that was trained with anime screenshots.
i dont have the prompts used for any of these images but, the prompt was pretty basic, i dont use any embeddings, so it was like: score_9, score_7_up, source_anime, [other quality tags], <REVY-LORA>, BREAK, rest of the prompt. for the negative: score_5_up, score_4, source_pony.
Also why i dont use score_8_up, it always gives me semi realistic or 3d results, i dont know if it is just for me but when i dont use it, it gives me more flat coloring results !
@Taz_ia I see. Thanks for the info.
@Taz_ia You should make an article or something on this site detailing what you use, since you use a mix of models and loras it seems like it would take quite some effort to replicate the previews. Also Civitai gives you the option to add prompt and settings details in the images when you upload them, that would be very helpful because even a slight difference in something like cfg can change a lot. Great loras btw, some of the highest quality I've seen
what is you generating data? i cant replicate same look as yours, tried different pony models, screencap loras and weights but cant make it look like yours. do you use adetailer?
I do use a mix i made of multiple models and loras, adetailer is good but i don't use it, i just go into img2img, me personally i think its better.
Also if you are trying to merge a model, i suggest merging with cartoonish models with a low weight, because normal pony models or anime pony models tend to give semi realistic or 3d results.
For the data on the images i don't have them because when i use photoshop, it deletes all the metadata.
@Taz_ia Is it really difficult to share the mix you use? Or you want to keep it for yourself?
Thanks for making this lovely lora!
FYI, I believe Revy's hair is actually supposed to be burgundy colored, not exactly purple, and her eyes are amber-brown, but still pretty great!
Mm, Revy-chan. ❤️
Yeah it would be great if the AI wasn't so simple & strict with captions.
You could give it a try, but i don't think "burgundy hair" would actually have any effect. Same thing for "amber-brown"
Thank you very much for liking my Lora !! :)
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.