This LoRA was trained on the VIPER RISE (GTB, GT1, V16) game series that I especially like because of the very unique artwork it had. After days of preparation, it's finally done.
Over 90 screencaps of input animations (from GIF sources) were used. Ranging from close up, to mid shots and full body drawings. After that all images were manually upscaled - and then refined (IMG2IMG) using an early version of this LoRA in Krita utilizing the Diffusion plugin.
It was also uncensored this way - and to my surprise this worked almost too good. Image errors (because of upscaling (4x ESRGAN) or poor input quality) were removed with Krita Diffusion, as well. All without straying away from the original artwork.
Tagging was done by hand using booru tags and special refinement on camera positioning, plus additional scene context.
This LoRA does (not) know any character names, but it works REALLY well with other character LoRAs from VIPER - or anything else I experimented with. Please be aware that using too many LoRAs will give you weird / overtrained outputs, so reduce the weights accordingly.
DPM++ 2M Karras and Euler samplers work best. Using the (8-step) Lighting LoRA, I usually go with 14-16 steps with a CFG of 2.5 to 3, Clip 2.
Description
Presumably the last version. I added in character names, which results in more adherence to the original material they're based on, but since this LoRA was - not - trained on concepts, but on a style (difference is, that I captioned every fine detail of the scene instead of leaving out the traits that describe a character specifically), the results are not always the character themselves, but an approximation.
Names / Keywords are:
Akira Honjouji
Makoto Takanami
Raika Grace (technically, she is not part of the VIPER RISE series)
Karin
Irear
Narsha
Saki
FAQ
Comments (14)
The default settings for this are ridiculously high. It sets the strength to 3.00 in the the generator. You have to manually type in 1.00 to get anything remotely usable.
Didn't notice that before. You mean the value after ":" (usually set to ":1")?
@redhome I'm talking about the slider in the on-site generator. The minimum I slide it to is 3.00 so I have to type in the input box '1.00'
To be honest I don't use the "on site" generator at all. So this is more a matter of CivitAI, I guess. It has not much to do with the LoRA. Or at least I cannot directly control the default values. Using ComfyUI / A1111 (Forge) everything is fine.
Hey man, been using this alot from before, since civits recent fix the lora is unusable because the slider doesnt go down 3 anymore, was wondering if you can do anything about it?
I tried editing the details for the 1.8 model. Please try again. :)
@redhome You're a legend! Just an fyi we cant go above 1, but that's ok for me since I always use 0.6 to 1 anyway, thank you very much!
@bmth1908 You're welcome! :) Going above "1" will introduce errors in most cases, but I changed this value too. Maybe it helps.
@redhome hey man back here again can you do what you did with 1.8, to the other versions aswell. Been using 1.5 and had to remix from old images to get a .8 setting.
I just updated this for all versions.
@redhome hey man Thanks I'm the same guy, this is my main account I'll probably bombard this with post, the results are awesome!
@karljeff1908 Thanks for the love. Maybe I finish my Viper RSR LoRA in the coming weeks. It's sitting on the shelf for a long time.
@redhome oh please do and maybe i can finally have sone decent Cala pictures haha I havent gotten lucky with finding decent settings that can make the existing loras work
Details
Files
Available On (3 platforms)
Same model published on other platforms. May have additional downloads or version variants.