v2.0:
权重需要比v1.0降低一些,也许0.8-0.9差不多
lora weight is suggested at 0.8-0.9 in the version
v1.0:
I cant promiss that it can work well.
效果不敢保证,但实现了小饥的lora从0到1的跨越。
We all know that there is spot on face of fami,but it wont appear if weight is low so just adjust weight when you cant find spot.
小饥脸上的斑在lora权重低的时候可能不会显现,此时可以稍微拉高权重。
Controlnet is adviced to use to control gesture firstly.
Here are sample pics from 20 different models.
Sample pics are all generated at weight 0.95.
Sampler :dpm++2m
res:768*768
tag:
pos:
(4k, RAW photo, best quality, masterpiece:1.2), (realistic, photo-realistic:1.37), ultra-detailed,photon mapping, radiosity, physically-based rendering,cinematic lighting, 1girl, upper body, looking at viewer, school uniform, basketball cap,
neg:
EasyNegative, paintings, sketches, (worst quality:2), (low quality:2), (normal quality:2), lowres, ((monochrome)), ((grayscale)),extra fingers,fewer fingers,strange fingers,bad hand,signature, watermark, username, blurry, bad feet,bad leg, duplicate, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands,mutated hands and fingers,bad-picture-chill-75v,badhandv4,
展示图片都是在权重0.95生成,这也是较为推荐的一个设置,上面是示例图片的tag,最好用controlnet先控制好姿势。
Description
v2.0 is not better than v1.0
choose your favorite one
FAQ
Comments (11)
로라 설정은 어떻게 해야하나요?
모델은무엇을 써여하나요
What is the source of the picture when you make Lora? The unique drawing style is so pretty.
manga
@Bruce_Leethanks:)
I'm sorry, but I can't find the information even if I search for hash, is it a self-made model? It's a shame that I can't understand the painting style is so prettyㅜㅜ
not self model,there are generated by common models
@Bruce_Lee I've been waiting for your answer, thank you so much for your reply. Then can I know at least the model name applied to the 5th and 12th pictures?(If possible, the first and third...)
@haramloveming351 12th is model pastel,others i forgot already
@Bruce_Lee Oops, I see. I appreciate your pointing it. Then, can you tell me the model names that you usually use when you draw illustrations?TT((lastly) Most of the models you use are so pretty
@haramloveming351 if u mean models i usually used,like doll,chillout,cetus,pastel,old fish,abyss and so on













