The first version of this LORA was trained on 1200 images from the game "Being A Dik". The game is an interactive experience created by DrPinkCake that features content similar to movies such as American Pie.
All 1200 images come from episode 2,8 and 9. A following Lora will probably contain more episodes. This LORA makes generated images look like Daz 3D renders, especially in the style of Being A Dik. It is prone to NSFW images because of its dataset, but can easily produce SFW images when using a good prompt. It can do characters (both male and female), enviroments and objects to a certain extent.
In theory, this LORA could be used to fix poorly rendered DAZ images or even create a Visual Novel from the ground up.
This is only Version 1. More will follow.
This Lora can be combined with specific character Lora's to specify characters from different Visual Novels. This LORA works very well with the Badik maingirl LORA's created by WaffleAbyss: https://civarchive.com/user/WaffleAbyss
Isabella: https://civarchive.com/models/107512/maya-bailey-being-a-dik
Jill: https://civarchive.com/models/107547/jill-royce-being-a-dik
Josy: https://civarchive.com/models/107501/josy-taylor-being-a-dik
Maya: https://civarchive.com/models/107512/maya-bailey-being-a-dik
Sage: https://civarchive.com/models/107562/sage-morgan-burke-being-a-dik
Description
V1 is trained on ±1200 images. It does require Hi-res fix to look decent. ADetailer is optional but gives even better results. The model should work without a trigger word, but works better when you do use it. I have only tested it on more realistic checkpoints, so I do not know what this LORA will look like when used with a 2D Checkpoint.