Add a ❤️ to receive future updates.
The model has been carefully optimized for stability and reliability, and its checkpoint is regularly updated to ensure that it incorporates the latest advances in deep learning research. This means that you can trust the model to produce accurate results time and time again, making it an ideal tool for artists, animators, and anime enthusiasts alike.
So if you're looking for a powerful and dependable anime model to help you create stunning art and animations, look no further than the "Anime Model OtakuReliableEnable (AMORE)" anime model.
模型不含垃圾数据无效精度,请放心下载!
插播:
如果想要更好的q版人物图,可以看看这个模型:
QteaMix 通用Q版模型 - gamma | Stable Diffusion Checkpoint | Civitai
2023.4.26
V1与ICE版本更新,这两个版本没有任何的承接关系,V1基于copper模型微调,ICE基于Anything2.1H微调,二者并无关系。
因不可抗力,目前V1版本已经下架。
2023.5.8
channel版本更新
依旧是基于Anything2.1H微调,训练集采用2233张由通配符生成的图片:
KawaICE Channel是目前最为好用的幼态体型角色特化模型,欢迎尝试!
本模型为幼态特化模型,但并非不能出成女等体型角色,如果达不到想要的效果,请加入“small breast”等提示词。
It is not impossible to produce adult female and other body types. If the desired effect is not achieved, please add "small breast" and other prompt words.
本模型在Anything2.1H的基础上进行训练,修改了如“Small chest”、“kawaii”等与模型主题相关提示词的权重。
This model was adjusted using the full Anything2.1N model to modify the weights of prompts such as "Small chest", "kawaii", etc. related to the model topic.
【PS:】请不要添加“masterpiece”等起手式,除非画面出现各种问题
Please do not add "masterpiece" or other prompt words
本模型的VAE没有任何问题,无需使用其他VAE
There is no problem to VAE of this model, there is no need to use other VAE
【两个版本】
ICE:最推荐下载的版本,适合正常使用
Merge:适合融合的版本,如果你想要把这个模型融合到其他模型里面,推荐下载这个版本
当然了,跟所有的模型都一样,这个说明放在这里只是为了叠甲,该怎么玩怎么玩就是了。
Of course, as with all models, this instruction is just there in case something goes wrong.
你可以随意将本模型融合到其他地方,但如果你共享该融合模型,别忘了标注一下
We allow you to merge with another model, but if you share that merge model, don't forget to add me to the credits.
除此之外允许任何人复制和修改模型,但是请遵守CreativeML Open RAIL-这里是M.CreativeML Open RAIL-M相关内容:
Anyone else is allowed to copy and modify the model, but please comply with CreativeML Open RAIL-M.You can learn more about the CreativeML Open RAIL-M here:
模型可以和其他模型一样随意使用,但是请遵守所在地区的法律法规,以免造成麻烦(我们并不负责)
The model can be used as freely as other models, but please abide by the local laws and regulations, so as not to cause trouble (we are not responsible for it).
Automatic1111: SD-WebUI
Danbooru:打标器,帮助标准训练集
通配符Wildcards:没有这个东西我就不可能整出这样的训练集
秋叶甜品店:想法!
QteaMix:通过研究这个模型得到的思路!
Legacy Toon Model
See Credits Below.
I enjoyed experimenting with this mix and didn't like that it was nuked off the internet leaving me with only a local copy for backup.
Already pruned / fp16 / safetensor, did not modify this at all.
Prompting Info
Pretty simple to use
Clip 1 & Clip 2 both work, try them out
If you put in prompts like 'illustrated' or 'high def' you can push this model to be more detailed.. but that sort of defeats the point
I don't show this in any of the samples.. but because of the super clean line work and color palette, you can mix this with a lot of different LORA's and get some cool stuff.
Recommended Negative Prompt (Not Strict - Experiment):
worst quality, low quality:1.4), easynegative, bad-hands-5,
Credits
TheFoodMage
the_toon_goon
Custom merge made specifically to produce angels. Each version has strengths and weaknesses.
AngelMix -- My first mix. Makes wings probably around 90% of the time. Most your angels will come out good. Poor range for much else in my opinion. It can do other things, but not to the same standard.
AngelMix2 -- Wings don't come out perfectly quite as often, but it has a significantly higher range. Can make landscapes, animals, buildings, 3d, hardcore nsfw content, and of course angels.
AngelMix(Anime) -- Merge of 8 models. It can do pretty much everything. Has a problem with hands. If you only download one of my models, then this is the one I recommend.
Follow https://twitter.com/AngelTycoon for updates. Make sure to tag me with your creations!
Recommended use EasyNegative for negative prompt and built-in VAE.
Created with a simple merge from CoffeeMix_V20 and RealdosMix
I wanted to try something like this...
The result is not bad 🙂
////////////////////////////////////////////////////////////////////////////////////
Recommended Settings - VAE is already included
Clip skip : 2
Hires. fix : R-ESRGAN 4x+ Anime6B / Upscale by 1.5+ / Hires steps 14 / Denoising strength: 0.4
Adetailer / Just select face_yolo8n.pt
Sampler : DPM++ 2M Karras / DPM++ SDE Karras / UniPC
CFG : 8 Steps : 25
Prompts
(best quality, masterpiece)
Neg : (worst quality, low quality)
////////////////////////////////////////////////////////////////////////////////////
Do you like my work? check out my profile and see what else!
And A cup of coffee would be nice! 😉
////////////////////////////////////////////////////////////////////////////////////
For anything other than general personal use, please be sure to contact me
You are solely responsible for any legal liability resulting from unethical use of this model
This model can create delicate and beautiful 2D female characters.
newmarsmix
My Fans Lora : fanslora.com/miaou
Warning: Add "nude, naked" to the negative prompt if want to avoid NSFW.
Automatic1111 user :
i use generally :
Sampling method : DPM++ 2M Karras
Sampling steps : 25
CFG scale : 8
Stablegram
my fanslora : fanslora.com/miaou
US/European instagram woman style
Warning: Add "nude, naked" to the negative prompt if want to avoid NSFW.
(Chilloutmix base) Trained with EveryDream2, +70 000 pictures like :
Currently working on a new version with better accuracy and consistency. By the way, have fun and don't hesitate to share your images !
A marge that goese along with CGI/Digital Arts. photorealistic prompt are adviced.
This CKPT is from a quick project test fly which has a commercial pedestrain in it, so it's not that really powerful kind of model, but still interesting, and might be helpful if you are working on commercial projects.
(plz download the posts for more info about prompts. It seem I put a lots of wrong information into the neg prompts by a mistake....but still, the img is there anyway...)
below are without control net.
below are with control net.(0.65 of depth, just for stablising the space structure.)
Have fun
Contect me on Wechat: PJSJ2019
or Email: jlmaoju@outlook.com
Hello everyone, I am an architect and museum exhibition designer,you can call me 'Moogone', I am very frustrated that it is difficult to find AI models for public space interior design online. Therefore, I have decided to take matters into my own hands. I will be publishing various types of interior design models for public spaces on my Civitai account. These may include museums, planning halls, corporate management spaces, temporary exhibitions, game exhibitions, and other categories. Please follow me to avoid missing out. My wechat number is 'Moogone', if you have any suggestion please add my wechat.
大家好,我是一名建筑及博物馆展陈设计师。我很苦恼为什么网络上很难找到公共空间室内设计的AI模型,所以我打算亲自动手了。以后我会在我的Civitai账号上发布各种各样类型的室内公共空间设计模型。可能会涵盖博物馆、规划馆、企业管理、临展、游戏展览等各种各样的类别。大家请关注我以防失联。我的微信名叫做 'Moogone',有反馈或者交流可以随时加我,但是请备注civitai哈~
这是一个关于图书馆的dreambooth模型,因为我最近在做图书馆室内设计,所以整合了这个模型用于项目前期生成图书馆的一些效果图,我用了30多张选定的图书馆意向图,这些意向图精挑细选,涵盖的方面比较多,所以我为此对每一张图片编辑了详细的.txt描述文档,以此控制模型不过度发散,但是我同时也在每一个文本中放了一个token:ohwx,所以当你要调用这些数据的时候最好先在开头输入ohwx.
ohwx
library
A more curvy body with calming Details, But still a Petite
So many improvements!
////////////////////////////////////////////////////////////////////////////////////
Recommended Settings - VAE is already included
Clip skip : 2
Hires. fix : R-ESRGAN 4x+ Anime6B / Upscale by 1.5+ / Hires steps 14 / Denoising strength: 0.4
Adetailer / Just select face_yolo8n.pt
Sampler : DPM++ 2M Karras / DPM++ SDE Karras / UniPC
CFG : 8 Steps : 25
Prompts
(best quality, masterpiece)
Neg : (worst quality, low quality)
////////////////////////////////////////////////////////////////////////////////////
Do you like my work? check out my profile and see what else!
And A cup of coffee would be nice! 😉
////////////////////////////////////////////////////////////////////////////////////
For anything other than general personal use, please be sure to contact me
You are solely responsible for any legal liability resulting from unethical use of this model
Creates (hopefully) Crash from Quake 3 or Doomgirl from the Doom games.
Quake 3 Crash
Doomgirl
Ultimate Doom
R is more realistic, B is basic, more cartoonish
I plan to refine the model with two tracks in the future
////////////////////////////////////////////////////////////////////////////////////
Recommended Settings - VAE is already included
Clip skip : 2
Hires. fix : R-ESRGAN 4x+ Anime6B / Upscale by 1.5+ / Hires steps 14 / Denoising strength: 0.4
Adetailer / Just select face_yolo8n.pt
Sampler : DPM++ 2M Karras / DPM++ SDE Karras / UniPC
CFG : 8 Steps : 25
Prompts
(best quality, masterpiece)
Neg : (worst quality, low quality)
////////////////////////////////////////////////////////////////////////////////////
Do you like my work? check out my profile and see what else!
And A cup of coffee would be nice! 😉
////////////////////////////////////////////////////////////////////////////////////
For anything other than general personal use, please be sure to contact me
You are solely responsible for any legal liability resulting from unethical use of this model
这是一个偏平涂风格的二次元模型,建议使用时tag尽量选择单词或短语,务必写出明确存在并容易描述的词,不要写含糊不清的词或句子,比如你想要一个明亮的眼睛,就直接写“水灵灵的眼睛”,而不是“漂亮的眼睛”
采样方法:开启高清修复时建议DPM++ 2S a Karras,不开启时选择DDIM,
采样迭代步数(Steps):推荐45/100,小图45左右,大图100
CFG:10~16
高清修复设置:
潜变量 (最近邻-整数)
高清修复采样次数:
20~52
重绘幅度(Denoising):
57
放大倍率:
2
VAE建议选择橘子最新的版本
一个通用的起手式:extremely detailed CG unity 8k wallpaper,illustration,lens 135mm,masterpiece, best quality,young girl,naughty face,aquagradient eyes,
This was initially an attempt at making Brainime v3 but things got a bit out of control and while it's definitely capable of doing anime, there's something just so funny about the way it render humans...
Enjoy, for the laughs.
大家好,我是chosen鲷,这次带来的是我自己常用来画碧蓝航线人物的模型。
首先说一下,这并不是我训练而来的模型,而是我用0.4 camellia大模型和0.4cetus还有0.2的meina融合而来的。
效果还蛮不错的,和各路vae相性都很好,例如pastel kl-f8 8400之类的,它带有很鲜明的韩风厚涂的感觉,但是脸又带有一丝meina模型特有的扁平化感觉,形成了一种很有趣的风格,尤其适合跑碧蓝航线的各种舰娘,和那些lora相性都很棒。
关于提示词,正面常用的是masterpiece , (8k, RAW photo, best quality, masterpiece:1.2), ultra-detailed,这些
负面(worst quality:2), (low quality:2), (normal quality:2), lowres, normal quality, ((monochrome)), ((grayscale)), skin spots, acnes, skin blemishes, age spot, (outdoor:1.6), manboobs, backlight,(ugly:1.331), (duplicate:1.331), (morbid:1.21), (mutilated:1.21), (tranny:1.331), mutated hands, (poorly drawn hands:1.331), blurry, (bad anatomy:1.21), (bad proportions:1.331), extra limbs, (disfigured:1.331), (more than 2 nipples:1.331), (missing arms:1.331), (extra legs:1.331), (fused fingers:1.61051), (too many fingers:1.61051), (unclear eyes:1.331), bad hands, missing fingers, extra digit, (futa:1.1), bad body,
采样方式,sde 或者dpm++2mkarras表现很不错,其他的采样方式也各具特色,推荐大家多试试,步数32即可,不用太高,然后要点上高分辨率重绘,重绘幅度推荐0.5,重绘方式R-ESRGAN 4x+ Anime6B,重绘步数12。我常用的分辨率格式是800*400,这样重绘放大一倍就是1600*800,清晰度已经蛮不错了。
大家好,我是chosen鲷,这是我常用来画动画同人图的一个模型,风格比较清新自然,而且兼容性很棒,分享给大家。
这是一个融合模型,他是用0.5的colorbox和0.5的chosen-mix融合而来。
风格比较清新,明亮,同时具备了chosen-mix的兼容性和colorbox优秀的色彩表现,从r-18到文艺小清新都能有很棒的表现。
推荐tag,正面:(Masterpiece:1.2, high quality)
负面:(worst quality:1.4), (low quality:1.4), (normal quality:1.4), lowres,EasyNegative,text, badhandsv5-neg
采样方式dpm++2mkarras 步数32步,记得点上高分辨率重绘,重绘方式用Latent,重绘幅度0.5~0.7,尺寸我常用的是800*400,放大一倍之后就是1600*800
Модель включает в себя лучшее, на мой взгляд, сочетание реалистичных стилей. Она соединила в себе более 15 различных моделей в четко выверенных пропорциях, что позволяет получать гипер реалистичные портреты невероятного качества. Модель натренирована на многих знаменитостях, рисует их лучше, чем любая Лора или Ликорис. Также содержит реалистичных nsfw контент.
The model includes the best, in my opinion, combination of realistic styles. It combines more than 15 different models in well-adjusted proportions, which allows you to get hyper realistic portraits of incredible quality. The model is trained on many celebrities, draws them better than any LoRaor Lycoris. Also contains realistic nsfw content.
I have created a new merger model that enables more realistic representation, thanks to the use of BRAV5.
Commercial use is strictly prohibited in v1.0. However, in v2.0, the terms and conditions of chilled_remix_v1vae which are described below, will be inherited and allow for commercial use.
OK:Use the model without crediting the creator
OK:Sell images they generate
OK:Run on services that generate images for money
OK:Share merges using this model
OK:Sell this model or merges using this model
NG:Have different permissions when sharing merges
0.7(0.6(braBeautifulRealistic_brav5) + 0.4(chilled_remix_v1vae)) + 0.3(V08_V08)
ng_deepnegative_v1_75t,badhandv4,paintings,sketches, (worst quality:2), (low quality:2), (normal quality:2), lowres, normal quality, ((monochrome)), ((grayscale)),logo,text
YabaL_Mix True25D V1.0
- I have added a lightweight version with fp16 (경량화 모델 추가)
- Recommended settings (추천 세팅 값)
- Sampler : DPM++ 2M Karras
- Sampling Steps : 30
- Upscaler : R-ESRGAN 4x+ Anime6B
- Hires Stpes : 15 - 20
- Denoise : 0.45 - 0.55
- CFG Scale : 6 - 10
- Used DDtailer : all settings are basic but Mask Blur is : 8 (DDtailer 블러 값 8)
- VAE : vae-ft-mse-840000-ema-pruned
- You can use to NSFW (NSFW 가능)
- "Cinematic" Prompt is working very well. I recommeded it! ex) (Cinmatic:0.4),
- Lora : I didn't used for Sample Imagies. (로라 사용 하지 않음)
Sample Image Prompt :
(masterpiece:1.2), (best quality:1.2), night, (looking at viewer), bright skin, {full body|upper body|cowboy shot},
outdoors, fringe, long hair, (wavy hair),
Negative Prompt :
worst quality, low quality, watermark, text, error, blurry, jpeg artifacts, worst quality, low quality, normal quality, jpeg artifacts, signature, username, artist name, bad anatomy,
EasyNegative, letterbox, tattoo, text, letterboxed, colored skin, umbrella, fog, crossed arms, crossed legs, moon, sun, white, censored, mosaic censoring, panties, bar censor,
arm support, navel, jeans, off shoulder, cleavage, naval, scar, hat, cap,
该模型是以novelai官方animesfw-latest模型202fcec0(旧版短哈希值)为底部模型融合而来
该模型使用难度可能会比较高,因为是以官方模型为基础所以会出现画面崩坏的情况,在其他模型可以使用的prompt可能需要重新调整,但是使用得当会有一些意想不到的效果
该模型可以不使用vae
fMix - one of my favorite
Prompt can be long or short, individual words or long sentences, the painting style will be affected.(<75 is better)
Commercial use? Ask me for the mixed model list and go ask their original author. (Just a suggestion)
The example is generated with wd-2 vae and "No xformer","use cpu all" (or webui-colab default)
Please share your amazing works here if you would like to :)
If you would like to see the recipe or more detail, feel free to ask.
since I use these amazing models to merge (which aren't belong to me), I won't use these models to make money (of course you can if you got the permissions from their original author)
if you really like my works and want to give me some kind of encouragement, consider click on the link to find out what it really is (didn't know what icrc is...?), I will be grateful(You could pick a different project/website if you like)
Attempted a large model in comic style, but the result seems mediocre
提示:男性眼睛会稍微有点问题,还有手,最好加上相应提示词、
推荐使用:高清修复、Euler a、重绘0.45。
2.5合并来源:
Akira Toriyama Style LoRA - toriyama_akira_3 | Stable Diffusion LoRA | Civitai
Pika's New Generation v.1.0 - v1.0 | Stable Diffusion Checkpoint | Civitai
Abara Heiki Sketches (Art Style) - gamma | Stable Diffusion LoRA | Civitai
Hand_Drown Style - v1.0 | Stable Diffusion LoRA | Civitai
Old Sketch Style - v1 | Stable Diffusion LoRA | Civitai
2.0合并来源:
Akira Toriyama Style LoRA - toriyama_akira_3 | Stable Diffusion LoRA | Civitai
Pika's New Generation v.1.0 - v1.0 | Stable Diffusion Checkpoint | Civitai
Abara Heiki Sketches (Art Style) - gamma | Stable Diffusion LoRA | Civitai
Hand_Drown Style - v1.0 | Stable Diffusion LoRA | Civitai
Akira Toriyama Style LoRA - toriyama_akira_3 | Stable Diffusion LoRA | Civitai
Pika's New Generation v.1.0 - v1.0 | Stable Diffusion Checkpoint | Civitai
sketch,
ThePinkDream_v01模型是基于v08,按照个人偏好的比例融合了GuoFeng3.2, fashionGirl_v52, chilloutmixss30_v30, iu_V35, koreanDollLikeness_v15.
通过微调模型使其处于真实以及2.5D之间较为不错的位置。
该模型在NSFW内容也有较为不错的表现。
如果出现对比度较高的情况,可以配合 clearvae_main 使用。
注:构建这个模型来帮助用户使用Stable Diffusion进行Ai绘画创作如图例所示的内容( 包括NSFW 类型)。但请注意,该模型并不是完美的,也不能保证其准确性和可靠性。作者不对该模型的使用结果承担任何责任,也不对使用该模型引起的任何后果负责。用户应该自行承担使用该模型的风险和后果,并在使用之前仔细阅读和理解本免责声明。
※2023/05/07
DefbeaCF3AR2D_mix_v2を追加しました。
数時間前に誤ったバージョンをアップロードしていました。現在アップロードされているものが最新です。
BRAをV4からV5に変更しました。
DefactaをDefacta3thに変更しました。
手の破綻が少なくなっている気がします。
DefbeaCF3AR2D_mix_v2 has been added.
We had uploaded the wrong version a few hours ago. The version uploaded now is the latest.
Changed BRA from V4 to V5.
Changed Defacta to Defacta3th.
I feel that the hand is less broken.
・Defacta3th
Defacta3th - v1.0 | Stable Diffusion Checkpoint | Civitai
※2023/05/04
DefbeaCF3AR2D_mixを追加しました。
顔に対してAnimelike_2Dというモデルを少量適用しました。
目のハイライトの入りが若干良い気がします。
現在Animelike_2Dがどこにあるかは不明です。
Added DefbeaCF3AR2D_mix.
Applied a small amount of the Animelike_2D model to the face.
I think the highlighting of the eyes is slightly better.
It is not known where Animelike_2D is currently located.
model_A | model_B | model_O | base_alpha | weight_name | weight_values
DefbeaCF3AR_mix_v1 | Animelike_2D_V2_Pruned_fp16 | DefbeaCF3AR2D_mix | 0 | 0,0,0,0.2,0.4,0,0,0,0,0,0,0,0,0,0,0,0,0,0.2,0.4,0,0.3,0,0,0
========
アニメ調のモデルです。
nsfwを目的として個人的にマージしていましたが、nsfw以外の出来もよかったので共有したいと思います。
人物の顔のパーツや背景の書き込みが好きです。
This is an animated style model.
I had personally merged it for nsfw purposes, but I would like to share it with you because I liked how it turned out outside of nsfw.
I like the facial parts of the person and the background writing.
以下のモデルをマージしました。
The following models were merged.
・Counterfeit-V3.0
Counterfeit-V3.0 - v3.0 | Stable Diffusion Checkpoint | Civitai
・AniReality-Mix
AniReality-Mix - v1 | Stable Diffusion Checkpoint | Civitai
・(1~6) Defacta
(1~6) Defacta - Defacta 6 | Stable Diffusion Checkpoint | Civitai
・BRA(Beautiful Realistic Asians) V4
BRA(Beautiful Realistic Asians) V4 - v4.0 | Stable Diffusion Checkpoint | Civitai
マージ比率はOrangeMixsを参考にしました。
Merge ratios are based on OrangeMixes.
WarriorMama777/OrangeMixs · Hugging Face
1.
model_A | model_B | model_O | base_alpha | weight_name | weight_values
16Defacta_defacta6 | braBeautifulRealistic_v40 | temp1 | 0 | 1,0.9,0.7,0.5,0.3,0.1,1,1,1,1,1,1,0,0,0,0,0,0,0,0.1,0.3,0.5,0.7,0.9,1
2.
Counterfeit-V3.0 * 0.5 + AniReality-Mix * 0.5 = temp2
3.
model_A | model_B | model_O | base_alpha | weight_name | weight_values
temp2 | temp1 | DefbeaCF3AR_mix_v1 | 0 | 0,0,0,0,0,0,0,0,0,0.3,0.1,0.3,0.3,0.3,0.2,0.1,0,0,0,0.3,0.3,0.2,0.3,0.4,0.5
this model is:
able to understand a wide range of topics, items, and concepts without additional Lora training
prompt weighted fairly even, most of the time you can just put prompt anywhere without worrying about weight and order
clear face from low res images, I can get a very clear face with 512x768 with text2image in 1 shot
limitations: not good at hyper-realistic, and you still need to take a few extra tries if you just rely on text2img and don't use control net or img2img( limited by SD1.5's capabilities)
というモデルです:
Loraのトレーニングを追加することなく、幅広いトピック、アイテム、コンセプトを理解できる。
プロンプトの重量はかなり均等で、ほとんどの場合、重量や順序を気にすることなく、プロンプトをどこにでも置くことができます。
低解像度の画像から鮮明な顔、512x768でtext2imageを使えば一発で鮮明な顔が撮れる
制限:ハイパーリアリスティックは苦手、コントロールネットやimg2imgを使わずtext2imgだけに頼る場合は、数回余分に試行する必要がある(SD1.5の機能による制限)。
该模型为:
能够理解广泛的主题、项目和概念,无需额外的Lora培训
提示的权重相当均匀,大部分时间你只要把提示放在任何地方,不用担心权重和顺序
从低分辨率图像中清晰的人脸,我可以从512x768得到一个非常清晰的脸,且通过text2img一发搞定
局限性:不擅长超现实,如果只是依靠text2img,不使用controlnet或img2img(受限于SD1.5的能力)的话,出好图任然依靠概率
check this prompt, its that all.
shiny sissy luxury latex dress for doll, subtle ambient glow, global illumination, lora:flat2:-1, Negative: EasyNegativeV2,badhandv4,worst quality,(noise:1.1),unnatural pose,improper proportions,facial asymmetry,disconnected elements,irrelevant background, stain, animal ears,
This merge model was created with the goal of reflecting the beauty of latex rubber enamel in a anime like illustration; V3 maintains that goal, but with less body shape and a prettier face. i also aimed to create a 2.5D model that is slightly closer to the illustration than before, making it easier to use.
Therefore, SSSSLLDDLLv3 can be used as a normal anime model. As shown in this image, even a very ordinary Japanese adult woman can be output in a cute way, confident in the expression of eyes.
standing, wide angle, (1girls:1.3), harem anime official art, seifuku clothes, school uniform, wide angle, glossy detailed pantyhose, large eyes big, cute beautiful eyes detailed pupils cute, circle white pupil cute, town park background, 35mm f1.4, iso100, raw, high quality,subtle ambient glow, raytracing, global illumination, ultimate quallity,
But you guys who see this image will think: "AYAYA I want to put this woman in a tight-fitting latex suit. No, I bet she has a rubber suit in her bag!" , Absolutely, this model was created to make your dream come true.
i have achieved fabric details like live-action while maintaining the cartoon face. i believe this is "art that is difficult to achieve with illustrations or 3DCG, but can only be expressed with AI models" !
sitting desk, wide angle, cowgirl position, ([precure:lovelive:0.1] face:1.1), (black serafuku:1.2), (fusion black tight school uniform and leotard with sissy frills:1.3), black cameltoe, pantyhose, black gloves, wet clothes latex, lip gloss, makeup, sweat, heavy breathing, (hime cut:0.8) ,black hair, (school classroom:1.2), evil smile, (vulgarity:1.2), school girl, blue eyes, skin glossy, shiny clothes, glistening fabric insanity detailed, subtle ambient glow, global illumination, lora:flat2:-0.4:NF,
Actually, V2, which was created with a different model than V1, was also partially published, but it suffered from low stability of the limbs. However, when we merged the GOD-LIKE LoRA, both the correction of the human body and the face of the character were improved. im decided to call this V3.
wide angle, missionary, ((kinky harajuku kei)), (sissy:1.4), (silky satin sissy long dress with many ruffles:1.3), (textile shading lolita fasion satin dress:1.3), (Glossy clothes luster:1.3), garter straps, (thighhighs:1.1), gothic latex corset, boots, (torpedo breasts:1.0), heavy makeup, eyeshadow, eyebrows hidden by hair, cheek blush, thick rouge lips, narrow waist, wide hips, large eyes big, circle white pupil anime cute, imperial palace room, subtle ambient glow, global illumination, ultimate quallity
The name of that LoRA is Big Hard Latex. As a side effect, I think the latex outfits are now designed closer to what is actually sold. It is still LATEX ......! LATEX solves everything ......!
Apart from that, we made the busts output with HUGE breasts larger.
((goddess)), (textile shading Navy Blue pvc suit with many ruffles), (angel wing, angel halo), gothic corset pvc, ((torpedo tits, wide hips, gigantic boobs, narrow waist, oppai loli)), Gray hair, seductive smile, in heat, , shiny clothes, glistening fabric insanity detailed, subtle ambient glow, global illumination,
If you are trying to produce a simple expression with this model, the following settings are recommended. Also, be sure to use Hires.
Sampler: DPM++ 2M (or DPM++ SDE) CFG scale: 4-6 Hires upscaler: 4x_RealisticRescaler_100000_G (or ESRGAN4x AnimeVideo) Steps:over 16 Hires Steps: 0 or half of Steps Hires Denoise: 0.6
Using this image as a basis, make various changes. Unmodified from below this image (fingers, etc.).
1girl, dynamic pose, dynamic angle, cinematic shadow effect, looking at viewer, cygames mecha_musume, Transformation by mechanization, demon lord girl, (kawaii:1.2), elf ears, demon horns, (metalic bodysuit:1.2), ([Mechanical exoskeleton:boned armor:0.5]:1.0), (glowing mechanical wings:1.2),((floating_armor)), female focus, beautiful detailed eyes, post apocalypse, subtle ambient glow, global illumination, ultimate quallity,
From here, upscaler was changed to latent and CFG was changed to 3. The clarity was reduced by the CFG effect, but the detail was improved. CFG 7-8 is recommended for a clear picture or for a high gloss effect. 9 or higher may result in poor rendering around the lips.
Also, by setting Sampler to UniPC, 3-5 CFG, and setting the Hires Steps to a value less than the Steps value, you can create a 3DCG or figure-like atmosphere. Conversely, the more you increase the value of Hires Steps, the more detailed the writing will be.
I was merging models while looking for other settings that would beautify the gloss, and in the last two weeks or so, two game changers in image representation have emerged. (so often the case in the AI industry LMAO)
One is the extension Self Attention Guidance. This can be done by setting the intensity from 0.75 to 1, which makes the image look like it has increased clarity. This is similar to how depth of field is adjusted with f-stop in photography. In the example, the simple background was left as is, but the background writing may be improved as well.
The other is iroiro-lora's flat2. This LoRA can be applied in the negative to increase detail/writing. In this example, in addition to the above SAG, we have given it a flat2:-1.
These two topics have synergy with each other. Other prompts such as costume wrinkles and hair detailing could be improved, tiled diffusion should also work, etc. All of these are my favorites! Please try them and experience the detail !!
highest quality, (trending on artstation), poster illustration, (2girls:1.2), (a High Quality Victorian Era cute girl), (luxurious intricate-harem dancer outfit with pelvic curtain:1.2), (cameltoe:1.2), anklet, (neck ornament and cleavage and covered nipples), bob cut, (intricate halter backless dress:1.2), sideswept hair, pixie cut, (see-through_dress:1.2), (beautiful face), (sexy eyes), detailed face, detailed eyes, slight smile, glitter many particles, lora:flat2:-1,
[EXTRA] I realized after writing these that CFG1.5 is the practical level only for UniPC. You are not hearing me wrong, it is CFG1.5, use it with SAG and flat2.
Finally, any liability arising from the use of this model and the media generated by this model is solely that of the user.
We have not tested whether it can be generated for NSFW media. Nudity, etc., ¯\(ツ)/¯ is often detected by AI in each publication, so please be aware of regulations and the laws of your country when using this service.
But even with less skin tone, you and your proclivities can create sexy, and masterpiece artwork.
Only on the LATEX.
standing, wide angle, (1girl:1.0), corruption, (dark persona magical girl:1.2), colorful hair, colorful clothes, dynamic poses, wide angle, latex gloves, highleg playboy bunny suit, showgirl skirt, glossy detailed pantyhose, large eyes big, cute beautiful eyes, effect in dark background, high quality,subtle ambient glow, raytracing, global illumination, ultimate quallity, lora:flat2:-0.6, lora:wowifier_v2:0.5, lora:Villagesuccubus_v30:0.8:NCNF
more infomation and v1 misc, check huggingface.
latex
ReaDiff is a realistic Stable Diffusion model that combines 3 other realistic SD models into 1 merged checkpoint. It can produce high-quality and realistic results for various scenes and subjects.
饺子皮是工场做的可是我答应质量很好。比手工还好吃!
A stylized anime model. This model is the result of attempts to get the textures from MouseyMix, a model that I'm very fond of, on characters with more realistic body proportions. The name 'Gyoza' part of 'Store Bought Gyoza' comes from a shortening of "Gyokai" + "Zankuro", two of the artists MouseyMix is trained on. The Store Bought comes from the fact that the 'bones' of this model, i.e. the primary compositional element is SD-Silicon, or to be more specific Silicon-29. SD-Silicon is made using autoMergeBlockWeight, an automated block merger, hence the dumpling is "made by machines". In other words, 工厂做的,商店买的。
The strengths of this model is that you get a very soft illustrative style model with bright texturing but with very powerful scenery and landscapes. V3 is the least overtuned of all the GyozaMix models, it recognizes fantasy, scifi, and normal settings. Though generally it does want to generate very grandiose landscapes so don't use it if you want something understated.
Everything is on this HuggingFace repository, along with the recipes. I go over all my decision making steps there. If you have anything to ask, don't be hesitant to leave a comment. I believe in an open-source approach to Stable Diffusion model mixing so I'll answer any questions you have on the process. It uses Block Weight Merging pretty extensively, so I would recommend you read up on UNet Blocks as a starting point.
But if you want a general rule of thumb, basically what we're doing is splitting the model mixing process into two parts: a picture composition/landscape arrangement component and a textural component. The former we can make by reverse-cosine Block Weight Merging any sort of highly detailed anime model against Silicon-29. I like to use dpep4 variants. Then, after we tune up the textual component to get the exact sort of texture we want, we cosine Block Weight Merge the textual component against our compositional component. After that, finetune for taste and we're ready.
皮子是买的可是馅子是新鲜的。
Generally speaking you will want to generate at a smaller size and then use some sort of General Adversarial Network based upscaler at a ratio of at, with a denoising ratio of around 0.3 to 0.4. This model can handle Latent upscalers but the result of that is less predictable. If you are using any sort of DPM++ sampler with Latent Upscalers, you need to upscale at a denoise ratio of 0.5 to 0.6, otherwise you image won't diffuse enough during the upscaling process.
Any sort of Negative prompt should work more or less fine. I like using the EasyNegative embedding because that usually adds a fair bit of detail into it, but as long as you prompt (worst quality, low quality:1.4) as a negative. Because of the training data, there's a good chance that you'll inherit some watermarks or signatures. If that bothers you, make sure to make that a negative too.
I can't really stop you but it's the wishes of a lot of people whose work went into this that you not monetize their stuff, and since this is a merge of their models, I want to pass that on to you. You can do whatever you want with the model once you download it, but please keep their sentiments in mind.
Odyssey is an expanded re-work of Ivory. It is a very versatile and competent model.
Semi-realistic anime style with good contrast and bright colors.
Some assets attained from MidJourney enable this model to produce fantastic landscapes/objects as well.
Works well with LORA/textual inversions.
For best results, generate at (512,768) or (768,512) resolutions.
NSFW capable although far from a strictly hentai model.
All sample images generated using Euler A sampling method at 20 steps.
Upscaled using Foolhardy Remacri at 0.3 denoising strength.
Images are not cherry picked to show true performance.
My own model mix, its a mix of a lot of models, naming started with 1_mix_a, 1_mix_b etc...
Latest Version 1_mix_s7
My focus leans more towards realism/fantasy, but of course it all depends on the prompt.
Be warned the model is mixed with nsfw models.
Use with standard VAE 'vae-ft-mse'
Hint for more realism add 'anime,manga,drawing, 3d render' to negatives.
One of my favorite merged models. Figured I'd share it.
I can't recall what I merged it with (sorry) but It does include Kenshi7thMix and also CSR that I like to use. Enjoy!
All preview images use this model
A few Prompts I like to use with it are:
Traditional Media
Thick lineart
Colorful
Outline
Sticker
(Gradients)
Dynamic
Vivid
Various colors
Platform
Art Deco
Splash art
Candy/Sweets/Treats/Delicious/Extravagant/Oversized ____/
by fitCorder
This checkpoint is a fine tune focuses on female fitness and bodybuilding. Keywords are not required. Works well with or without the Lora's i've uploaded.
Mogeko
Okegom
Funamusea
This model is fine-tuned on nearly 300 images of animal-human hybrids generated with MidJourney.
Fine-tuning was done at a resolution of 1024x1024 which allows it to natively produce images of that size without distortions or requiring Hires. fix
Since the training images were made with MidJourney, they were not explicit images and so you are unlikely to get those results when using the hybrid keywords (I have not gotten them in testing), but I cannot make assurances since it was finetuned from the RealisticVision 2.0 model located at: https://civitai.com/models/4201/realistic-vision-v20
Sampling Method: Euler
Sampling Steps: 16-50 (images shown use 16-32 steps)
the keyword "hybrid" is used for the anthropomorphized characters and in training it was formatted as something like: "[animal type]-[human type] hybrid"
"a photo of a fox-woman hybrid"
"a photo of a deer-man hybrid as a lumber-jack"
"a photo of a hedgehog-girl as a student"
In training the "a photo of" prefix was used with the photo-realistic images although many were in other styles so this is only useful if you want something more realistic. I also put things such as "cartoon, 3d model" in the negative prompting to get more realism.
Generic/unspecified animals | Badgers | Cats
Birds of various types | chameleons | cheetahs
deer | dogs | elephants
foxes | frogs | goats
hedgehogs | hyenas | lizards
llamas | monkeys | mice
rats | otters | owls
pandas (including red pandas) | pigs | bears
rabbits | raccoons | rhinos
rodents | seals | sheep
skunks | squirrels | turtles
and walruses
hybrid
OpenGen
OpenGen is a continuation from Protogen model, this model was handcrafted by a selection from what I personally consider the best models currently available, the licensing continues its respective formal bindings from its previous merges.
Adding images to gallery for better viewing
Ce modèle a été entraîné avec ~15-20 images haute résolution du sujet trouvé sur le web et le modèle v1-5-pruned-emaonly.ckpt comme base en utilisant le Colab Dreambooth de Shivam à la fin d'octobre 2022.
À la base un fichier checkpoint, il a seulement été converti en safetensors avec l'option Checkpoint Merger de l'interface Automatic1111. FP16 et Baked-in VAE.
This is a model that was trained with ~15-20 high resolution images of the subject found on the web and v1-5-pruned-emaonly.ckpt model as base using Shivam's Dreambooth colab in late October 2022.
Initially a checkpoint file, it was only converted to safetensors using the Checkpoint Merger option in Automatic1111. FP16 and Baked-in VAE.
***
Shivam Shrirao's colab:
Baked-in VAE:
vae-ft-mse-840000-ema-pruned.ckpt
negative textual inversions:
https://civitai.com/models/7808/easynegative
https://huggingface.co/nick-x-hacker/bad-artist
https://civitai.com/models/17083/bad-picture-negative-embedding-for-chilloutmix
extension:
https://github.com/Bing-su/adetailer
***
Ce modèle fourni par jeanguylafleur et mis à disposition exclusivement sur civitai.com a été créé strictement à des fins ludiques et éducatives.
Il n'est pas destiné à être utilisé de manière malveillante.
Ni l'auteur de ce modèle ni civitai.com ne pourront être tenus responsables d'une mauvaise utilisation ou d'une mauvaise application du modèle.
L'utilisateur assume la pleine responsabilité de toute utilisation ou conséquence qui pourraient découler de l'utilisation du modèle et est strictement interdit de commercialiser ou de privatiser le modèle pour son usage à but lucratif.
Veuillez utiliser de manière responsable.
Notez qu'il n'y a aucun plan pour entraîner davantage le modèle ou développer une version LoRA.
This model provided by jeanguylafleur and made available exclusively on civitai.com was created strictly for fun and educational purposes.
It is not intended to be used in any nefarious manner.
Neither the author of this model nor civitai.com shall be held liable for any misuse or misapplication of the model.
The user assumes full responsibility for any use or consequences that may arise from the use of the model and is strictly prohibited from commercializing or privatizing the model for their own for-profit usage.
Please use responsibly.
Note that there are no plans to further train the model or develop a LoRA version.
pspp man
Sexy Asia girl, material from the Internet,
It doesn't perform well on 16:9 landscape images, Recommended resolution: 9:16 3:4 1:1