
V2.0
-
Few more pics added to training
V1.0
Always liked the contrast from grey and bright colors.
I've created a few images with AI and models that know the concept somewhat and trained the Lora on a few images.
It isn't where i really want it and needs more training, but for now it's good enough.
-
Works best with more cartoony/anime models for some reason ( i always train on Base XL )
-
Some models know the concept and produce almost the same effect, but this Lora makes everything a bit clearer and in some cases changes everything entirely.
-
Tends to produce a lot of concrete walls, because most of the training pictures are about paintings on concrete walls ( which is why i need to make a few more images with sculptures of different subjects and landscapes )
-
Needs sometimes a few negatives to get rid of unwanted things.
-
Neon is of course a know word and produces a lot of times neon lights or neon glowing stuff, if unwanted putting glow as a negatives helps.
Well, i guess that's all for now. Would like to see if someone can tame that thing as it is at the moment.
描述:
训练词语: conrtneo,neon,concrete
名称: - SDXL - conrtneo _neon-concrete _V1.0.safetensors
大小 (KB): 111752
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success