LoRA for creation of an original VTuber Model
(EN/日本語/中文)
|___________________________________________________________________________________________|
About this LoRA:
This LoRA is for those, who want to create their own VTuber model, but have no artistic skills in illustrating one by their own.
For this purpose, the LoRA has been trained on images with white background and a strait standing character to ease the editing process afterwards.
___________
>Trained with >CouterfeitV3.0<
___________
>How to use it:
Personal checkpoint recommendation:
___________
Hard to use on:
AbyssOrangeMix
AnythingV3/V5
AnyLora
cetusMix
___________
IMPORTANT: I highly recommend using the OpenPose extension for ControlNet.
___________
-->Prompts: (simple background,white background:1.5), full body, standing, looking at viewer, solo,
-->Negative Prompts: cut out, cut head,(outpaint:1.5),extra leg, short legs, fat legs,
___________
-->LoRA strength: 1
-->Highres fix: Latent
-->Sampler: DPM++ 2M Karras
-->CFG scale: 7~11
________________________________________________________________________________________
Prompt example:
___________
Prompt:
masterpiece, best quality,1girl, full body,solo,(simple background,white background:1.5),standing strait, looking at viewer,open mouth, laugh,military uniform, pants, black uniform, gold patterns, hat, <lora:GoodHands-vanilla:1> vt <lora:VtuberV2-000007:1>,long legs,
___________
Negative:
1girl, EasyNegative, verybadimagenegative_v1.3,illustration, 3d, sepia, painting, cartoons, sketch, (worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, normal quality, ((monochrome)), ((grayscale)), cut out, cut head,(outpaint:1.5),jacket,extra leg, short legs, fat legs,
___________
*Important keywords are marked >BOLD<.
*重要なキーワードは >太字< で表記しています。
*重要的关键词被标记为 >粗体字<。
___________
________________________________________________________________________________________
このLoRAについて
このLoRAは、VTuberのモデルを作りたいが、自分でイラストを描く画力がない人のためのものです。
このため、LoRAは、その後の編集作業を容易にするために、背景が白で、キャラクターがまっすぐ立っている画像でトレーニングしています。
___________
>CouterfeitV3.0<でトレーニングしています。
___________
>使い方を教えてください:
個人的なチェックポイントのすすめ:
nullV22_nullV22+clearVAE
tmndmix + clearVAE
endlessmirai + clearVAE
notsoxjbmix1+clearVAE
___________
使いにくい:
AbyssOrangeMix
AnythingV3/V5
AnyLora
cetusMix
___________
重要:ControlNet用のOpenPoseエクステンションを使用することを強くお勧めします。
___________
-->プロンプトを表示します: (simple background,white background:1.5), full body, standing, looking at viewer, solo,
-->ネガティブプロンプト:cut out, cut head,(outpaint:1.5),extra leg, short legs, fat legs,
___________
-->LoRA strength: 1
-->ハイレゾ修正: Latent
-->Sampler DPM++ 2M Karras
-->CFGスケール:7〜11
________________________________________________________________________________________
关于这个LoRA:
这个LoRA是为那些想创建自己的VTuber模型,但没有自己的艺术技能来说明一个模型的人准备的。
为此,LoRA已被训练的图像与白色的背景和一个垂直站立的字符,以方便事后的编辑过程。
___________
>用>CouterfeitV3.0训练<
___________
>如何使用它:
个人检查点推荐:
nullV22_nullV22 + clearVAE
tmndmix + clearVAE
endlessmirai + clearVAE
notsoxjbmix1 + clearVAE
___________
难以使用:
AbyssOrangeMix
AnythingV3/V5
AnyLora
cetusMix
___________
重要提示:我强烈建议使用ControlNet的OpenPose扩展。
___________
-->提示语: (simple background,white background:1.5), full body, standing, looking at viewer, solo,
-->负面提示:cut out, cut head,(outpaint:1.5),extra leg, short legs, fat legs,
___________
-->LoRA强度:1
-->高像素修复: Latent
-->采样器: DPM++ 2M Karras
-->CFG规模:7~11