CamelliaMIx_2.5D - CivArchive (CivitAI Archive)

CamelliaMIx_2.5D - V2

NSFW

use vae : vae-ft-mse-840000-ema-pruned.safetensors, orangemix.vae.pt

V3 is closer to 3D than another version

Face shape and texture have been changed

▼Please check settings and prompts▼

I recommend using [dddetailer] - (Face correction extention)

Recommended setting

Sampling method : DPM++2M Karras, DPM++ SDE Karras, DPM++ 3M SDE Karras

Clip skip : 2

Hires steps : 13

Hires.fix upscaler : 4x-UltreaSharp

CFG Scale : 6~11

VAE : ft-mse-840000, orangemix.vae.pt

Recommended prompt

Prompt : [please refer to exmple images]

Negative : (worst quality, low quality:1.3)

Textual Inversion : negative_hand-neg, verybadimagenegative

Description

V2 is closer to 3D than V1
Shape of eyes and face have been changed

V2 is a slightly different style from V1

Merge model

CamelliaMix_2.5D

CamelliaMIx_NSFW_V1.1

AbyssOrangeMix2_hard

revAnimated_v11

Basil mix

sakimichanStyle_v15 LoRA

cuteGirlMix4_v10 LoRA

Process

0.5(AbyssOrangeMix2_hard) + 0.5(revAnimated_v11) = A

0.6(CamelliaMIx_NSFW_V1.1) + 0.4 (A) = B

0.8(B) + 0.2(Basil mix) = C

C + 0.1(sakimichanStyle_v15) + 0.1(cuteGirlMix4_v10) = D

0.5(CamelliaMix_2.5D) + 0.5(D) = CamelliaMIx_2.5D_V2

Recommended setting

Sampling method : DPM++ SDE Karras

Clip skip : 2

Hires steps : 13

Hires.fix upscaler : R-ESRGAN 4x+Anime6B

CFG Scale : 7~10

VAE : ft-mse-840000

Recommended prompt

Prompt : (masterpiece:1.2, best quality), (real picture, intricate details)

Negative : (worst quality, low quality:1.4), negative_hand-neg, verybadimagenegative

Textual Inversion : negative_hand-neg, verybadimagenegative

FAQ

Comments (14)

minacoApr 18, 2023· 2 reactions

CivitAI

welcome back~

MoreAIApr 18, 2023

CivitAI

What made you change your mind and come back after you tried to get rid of all the models you made?

Mods13

Author

Apr 19, 2023· 12 reactions

--This answer is in a summary. Because my English is not good. sorry I couldn't give you an exact answer--
I realized that there was no point in me making such an attempt because models are already spread across many communities. and i just thought it would be better for many people to use models.

dionisiosApr 24, 2023

@Mods13 hi wonderfull model, can i ask you, how i can make the color vivids looks a bit wahse out, thank you

MoreAIApr 24, 2023

@Mods13 Your English is fine, I'm just wondering why did you initially want to get rid of them? Were you not happy with their quality, or was it some other reason? I'm simply curious since you did such a good job with them.

xtianst80Aug 2, 2023

@dionisios apply "VAE" download one and place it in the vae folder, then go to settings to activate it.

YurojinApr 18, 2023· 6 reactions

CivitAI

Glad to see you back Mods13님

ninizi001Apr 24, 2023

CivitAI

欢迎回来，模型很棒！

406858515448Apr 24, 2023· 3 reactions

CivitAI

Be careful! When using the prompt ' very detailed face ', the picture will only focus on the upper body. And even if you use prompt like ' (full body:1.8) ', sometimes it still can't make full body picture.

Mods13

Author

Apr 24, 2023

When you write "detailed face" prompt, picture is focused on the face. So if you want to generait a full body picture, I recommend you don't use 'face' prompt.

If there is no "face" prompt, fullbody prompt is enough with 1.15 strength.

and If you want detailed face in a "full body pic", increase resolution or use ddetailer.

406858515448Apr 25, 2023

@Mods13 Thanks for you reply!

lxswApr 26, 2023· 2 reactions

CivitAI

I think the style of the face in this model is too strong. none of lora can change the face of character.

BeYonmeMay 7, 2023

CivitAI

为什么我不能在这个模型下分享图片？

Why I can not post image in gallery under this model?

IsekaiLanayayaMay 22, 2023