
LoRA for the Ultimate Bunny suit from the game First Descendant.
First time training a LoRA.
Trained from 44 screenshots from in-game. Had to crop the images but I wasn't consistent on the sizes. It turned out...okay, I guess. I definitely could have been more thorough in my prompts and chose better screenshots.
Full-body shots make the character a bit small relative to the canvas size due to how I cropped my screenshots. There's a bunch of things I wanna fix for a better version but I don't have enough buzz for more training, lol.
Recommended LoRA strength is 0.8
V1 Trigger Word: UltBunny
V1.5 Trigger Word: UltBunnySuit
Other Keywords: white bodysuit, helmet
I tested it with a few other colors and it seems to work. HiResFix and Adetailer (not recommended) almost forces a face to show up so try including facial traits in the negative prompt and add 'faceless' in the positive.
NOTE:
V1 Doesn't work with the site generator for some reason.
V1.5 Seems to work with site generation but I had to retrain the model (different img prompts used) so mileage may vary. I don't use the site gen so it lacks proper testing.
描述:
训练词语: UltBunny
名称: 653411_training_data.zip
大小 (KB): 76002
类型: Training Data
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: Ultimate_Bunny_Suit.safetensors
大小 (KB): 223102
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success