Lora model of Lavinia Whateley from Fate Grand Order. Preview images done with weight 0.9-1.
Needed to make something fast and low effort compared to previosu one :) Back when I was playing FGO I wished devs would at least make Lavinia into 1 star unit but they didn't T_T. At least now I can have more pics of her if I want.
Swimsuit is there but works right just once in a while. Mostly focused on her normal outfit. Maybe will retrained for it someday in future.
Normal version - lavinia whateley \(fate\), polka dot bow, black dress
Swimsuit version - lavinia whateley \(fate\), black bikini, polka dot bow, orange bow, see-through
Additional fun things to try:
constricted pupils - to make her pupils smaller. She usually gets wide eyes and small pupils even without it, but this brings it close to usual look. Can cause multiple pupils.
wavy mouth - works together with open mouth, closed mouth or smile to give her more awkward expression
nose blush - for extra blush
shaded face - for dramatic face shading
Description
FAQ
Comments (7)
Hi, I really like your lora creations. They look amazing! Could you please share some details about your workflow? I’m curious about how you do it. For example:
Where do you find the images that you use for lora? Do you search them on Google or somewhere else?
How many images do you usually use to make a lora?
What model do you use to train your lora?
What kind of GPU do you have and how long does it take to train your loras?
Do you use colab or kaggle for training ?
I hope you don’t mind me asking .I myself also want to make lora on different characters and things, so I’m looking for some guidance. Thank you.
I don't think I can be of much help because I still don't feel very knowledgable about this and I don't have an established method that I stick to.
Where do you find the images that you use for lora? Do you search them on Google or somewhere else?
I look through booru websites, Grabber program is useful.
How many images do you usually use to make a lora?
Between 25-500, it depends how many good images I can get and how many outfits character have. I probably average about 70 for simple models and 150 for bigger.
What model do you use to train your lora?
I train on NAI. I used to train on Grapefruit.
What kind of GPU do you have and how long does it take to train your loras?
I have laptop version of RTX 3070 Ti. Time varies greatly depending on settings.
Do you use colab or kaggle for training ?
I sometimes use colab for bigger models. I don't really like it but I feel it's about 2 times faster and I can play Genshin while it's training.
@Deto15 Thanks for the info (^∇^) .
Seems like i'll be using colab for training , my 1050ti will take forever to train.
BTW whats NAI model ? can't find it
@Ghost855 NAI is NovelAI, also called animefull-final-pruned. Not sure if I can give you links, you will have to find on your own.
@Deto15 Aaah ! thanks ! Found It 😊
Cute baka, undervalued fate girl, thanks for the lora
Amazing work! Thank you for giving this character much needed love ♥
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.