我第一次正经（？）制作的大模型，肢体会有一些问题，作为第二版已经优化。
This is the first time I've seriously (?) train a large model. There are some issues with its limbs, but it has been optimized in the second version.
Description
FAQ
Comments (5)
Looks pretty good.
Can you add generation parameters for your preview images? It makes it much easier to figure out how to use the model.
In fact, there should only be 4 images without parameters, you know, because these four triggered the TOS, just because they contain a word starting with L...
I uploaded 5 pictures, they contain picture information and seed information, you just need to add the tags of loli, little girl and adjust the weight to restore it, the information of the first picture cannot be uploaded, I will copy it to you
(masterpiece:1.2),(best quality:1.2),(illustration:1.2),[(star background:1.2)::10],(dynamic angle),(low contrast:1.2),(contrast colour:1.2),,(paint:1.3),(painting:1.3),(1girl:1.3),(disheveled hair:1.2), (white hair:1.3)/=(colorful highlight hair), red eyes, (detailed anima face:1.2), small breasts, bare shoulder, white dress、,(mini top hat),hair between eyes,braid,french braid,,manaphy,off-shoulder dress,,frilled capelet,,(colorful background:1.2),(surrounded by colorful splashes),colorful bubbles,(bright color),flat color,(floating:1.2),(rainbow color:1.3),(color ink:0.8), (ink splashing:0.8), (color splashing:1.2),
Negative prompt: text,username,logo,(low quality, worst quality:1.4), (bad anatomy), (inaccurate limb:1.2), bad composition, inaccurate eyes, extra digit, fewer digits, (extra arms:1.2)
Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 1101233828, Size: 512x768, Model hash: 42695aeac5, Model: komix-v3, Clip skip: 2, Version: v1.5.1
Then, if you have a good picture, please send me back the picture, thx