LyCORIS

This LyCORIS has been trained after carefully selecting 34 images. All photos have been enhanced to make the most of them. I hope you like it!

Any feedback would be highly appreciated!

From my experience it runs best with Edge Of Realism, pirsusEpicRealism, epicDiffusion ,RealisticVision, CyberRealistic, Deliberate checkpoints, among others, although I think it will be just fine in other SD 1.5 models as well.

CFG scale should be between 4-7.

LyCORIS weights should be between 1-1.4

Notice: This free work is intended for art and illustration purposes of emergent technology and not intended nor' trained for making illicit images, as the author I've taken care to not use any private data in the training process or designed it for such purposes. By downloading this model you are agreeing with the aforementioned intent and that you will not use my work to create / share images of a potentially immoral and / or illegal nature.

Mammon, Sin of Greed from the Seven Mortal Sins (Sin: Nanatsu no Taizai).

LoHa trained on NAI (anime). Requires the LyCORIS plugin.

Trigger words:

• mammon - For character. Optional tags: "wings". Eye color seems a bit inconsistent though, so adding "blue eyes" may help.

• mammonoutfit - For main outfit. Add "navel, cleavage, thighhighs, headband" for best results. The model is heavily biased towards this outfit, so triggers may not be necessary.

Experimental (undertrained) trigger words:

• mammonwitch - For witch outfit from mobile game. Add "witch hat, garter straps, thighhighs, halloween" for best results.

• mammonxmas - For Christmas outfit from mobile game. Add "santa hat, apron, clothing cutout, bell, red panties, thigh boots" for best results.

• mammonsummer - For summer outfit from mobile game. Add "blonde, bikini, sun hat, jewelery, see-through, thigh strap" for best results.

Welcome to Insta Thot Series Vol. 2! Realize your thot fantasy here!

--

If you love my work, you can buy me a coffee to help me stay awake and cope with life to create more lora and stuffs : ko-fi.com/tauron. It take less than 5 mins to like or leave a review while it takes me hours or even days to make Lora. Liking / Reviewing helps the thread alive.

Check vol.1 --> Tau's Insta Thot Vol. 1 [Fin] - 01. [Req] @Imazhai | Stable Diffusion LyCORIS | Civitai

Best to use with my mix: Tauron Hybrid Mix - TauHybrid_Anime_Final | Stable Diffusion Checkpoint | Civitai

--

DISCLAIMER:

I'm doing this for fun and not for commercial motive, My main purpose is to compose a fictional character using AI with the beauty of all persons above, please support them if you love their works. i do not support any indecent images generated of all persons in this model, I do not support selling generated images of persons in this model. please refrain to what they have been posting online.

All persons in this models are 21+ and not a minor, please DO NOT post explicit nudity, suggestive content, or pornographic content online using this model. Refer to CivitAi TOS for more details.

===

Request or Suggestion Rule:

1. No request of minor!

2. The requested should have public data (no private)

3. Should have more than 30 good quality images

4. Since all requests or suggestion are free, I will still need to like your request, otherwise its a work not a hobby for me.

5. Be a decent human being and leave a review if your requests is up It takes you only 5 mins while it takes me hours, electricity, and efforts to do your requests.

6. If you don't get what you want, do not review bomb, there could be other creators models or lora that can get better results than mine, use theirs instead, and leave a good review on them too, support any other creators for their hardworks!

NOTE : I will keep updating more actresses in versions update

===

Usage Guide

--

A. Lora Weight

• 0.1 ~ 0.4 -- is to combine multiple faces or features, or if you experience overbaking

• 0.5 ~ 0.7 -- is a balanced weight to get still accurate feature but with more flexiblity

• 0.8 ~ 1.0 -- is to create accurate representation of the features.

B. CFG Scale

• Lower CFG Scale means AI gets more creative and less overbaking problems.

• Higher CFG Scale means results get very accurate with training sets and more prone to overbaking problems.

• Dynamic CFG Threshold, using this method allows you to bump up quality and achieve photorealistic easily, but with lycoris i found out recently that this method actually lower success rate, so only use when you are composing with a Lora / model that works fine with this method.

C. Sampling Steps, High Res Fix & Upscaling

• I like to use sampling steps at 50, with high res step at 20, or 1/2 of sampling steps.

• I find that 2M Karras works better and faster than SDE Karras. with ERSGAN 4x+ method + ERSGAN 4x for Upscaling. This method actually gives quite good result without being too sharp or too plasticky

• Upscaling - GPFGAN @0.25, Codeformer @0.3 & Codeformer Weight @0.15 seems to yield superior results to me.

--

Combining with Attribute LoRa / LyCoris

You may need to do your own testing, but using character at weight -0.2 below attribute lora weight yields superior results, if for some reason one features persists not showing or overbake, you may want to lower the weight of the overbaked feature Lora, and then use emphasis () to boost the specific attribute you want.

DDetailer / ADetailer Script

If subject face is small / far away use: DDetailer Script with bbox anime face with default settings except denoiser at 0.2-0.3. You can combine this LoRa with a lot of other Style / Attire / Pose / Acts LoRa, Also works well with character LyCoris.

How to install this script? I kinda forget how because its a bit complicated beacuse you need to install stuff from venv. but a quick google search will land you somewhere in tutorial page.

Enjoy.

一组极具设计感的家居设计图片训练的模型。

搭配不同的底模可以解锁更多风格

A group of highly designed home design picture training model.

Pairing different bottoms can unlock more style

可定制商用模型

Commercial models can be customized

Patreon

Get early access to build and test build, be able to try all epochs and test them by yourself on Patreon or contact me for support on Discord

Dazzle AI

You can share images that you have generated on discord, discuss prompts and/or ask for help here in Dazzle AI Discord

You can also find my models on Dazzle AI wich provides users access to the Stable Diffusion models, enabling the generation of photorealistic images from simple text input in just ~6 sec

Recommended settings

• Make sure you are aware on the usage instructions of LORA

• Sampler: DPM++ SDE Karras (Recommended for best quality, you may try other samplers)

• CFG Scale : 5 to 10

• Lycoris weight for txt2img: reccomended 0.35

Those are a Lycoris, to use it you need to install the proper extension on Stable Diffusion.

Insutrctions

My advice is to start from the prompt of the posted images on civitai, or in any case use most of the keywords with 1.2-1.3 weight.

In general if you want a classic ironman suit you just need to use the keyword "wearing irongirl_suit" maybe with a weight if necessary.

If you want a mix like irongirl|ballet or irongirl|spider you should use both keywords, in this cases i have an important advice, play with weights! mostly when adding a particular color to the suit. I make you an example, most of the training ballet dress images are white, most of the training irongirl suit are red, so if you write "(wearing ballet dress with tutu, wearing white irongirl_suit:1.3)" SD will probably generate a normal white ballet dress, so in this case it's better to lower the weight for the ballet dress to have the mix you want, like "(wearing ballet dress with tutu:1.1), (wearing white irongirl_suit:1.3)". Opposite case if you want a red irongirl ballet dress suit.

Enjoy!

If you'd like to support my efforts and have access to early and bleeding edge builds, please think about joining my Patreon: https://www.patreon.com/betterCuteAsians

邪神ちゃんドロップキックのぺこらを出せます

Anything4.5を対象に768×768で学習させています

This Lora can output Pekola of Dropkick on My Devil!

training on 768x768 for Anything4.5

Patreon

Get early access to build and test build, be able to try all epochs and test them by yourself on Patreon or contact me for support on Discord

Dazzle AI

You can share images that you have generated on discord, discuss prompts and/or ask for help here in Dazzle AI Discord

You can also find my models on Dazzle AI wich provides users access to the Stable Diffusion models, enabling the generation of photorealistic images from simple text input in just ~6 sec

Insutrctions

Remember this is a Lycoris wich needs an additional extensions to be used.

My advice is to start with posted images prompt.

I really tryied to caption wearing_mask to avoid having the mask when wanted the face, but it didn't work, the generated girl comes out with a mask almost everytime even if only half of the training images have masks.

I tryied to put in negative prompt wearing mask and also tryied to put in positive with negative weight, really, nothing worked.

I hope you will have more luck then me while prompting or that you really like the mask.

Enjoy!

If you'd like to support my efforts and have access to early and bleeding edge builds, please think about joining my Patreon: https://www.patreon.com/betterCuteAsians

在此做出预告，我会将细分动作全部在本概念下更新，每一种姿势对应一个版本，请按需切换下载

Here is an update notice. I will update all subdivided actions under this lora. Each pose corresponds to a version. Please switch and download as needed.

版本之间不存在迭代差异，仅有姿势区别

There are no iteration differences between versions, only pose differences

目前已经更新第3个

The third one has been updated

好的手需要运气这件事我想大家应该不用我再强调了吧

Good hands need luck. I don’t think you need me to emphasize it.

下面是脱口秀时间，因为比较长还是一个梗，我就不翻译了

The following is the time for the talk show, as it is still a meme, I will not translate it anymore

总结一下就是，目前我试了几个已有的Pov missionary都是简单动作概念，我本来打算将Pov missionary的多细分概念合一，然而失败多次，决定拆分制作

The general content is that at present, I have tried several existing Pov missionary loras, all of which are simple action concepts.
I originally planned to combine the multiple subdivision concepts of Pov missionary into one, but failed many times, so I decided to split it.

我也想一个lora概念全包

是不是很大胆

一个lora概念全包

在Pov missionary这一个

就这个lora

就完全是这个lora   分概念全包（指包括什么leg grab、holding another's wrist、torso grab等等）

我最早收集数据集的时候

跟所有群友说的都是一个全包

难以想象

喔所有群友啊 

他们所有人都很兴奋（其实只有我）

一个全包一直说的是一个全包

我一直在做一个lora概念全包的准备

所以当时跟显卡说的时候

我只要你20天

或者只要你25天

10epoch到15epoch

显卡说为什么

我说很简单

我每天炼两遍

一锅lora两个小时

我每天炼两遍

上班炼一遍

下班炼一遍

我每天炼两遍

我炼十天

400多个lora我选一个

一个lora概念全包啊

这个Pov missionary就完了

我选一遍

那十天就够了

然后前面的十天打标

严格的打标

我说只要你10天咱就炼一超全的lora

就挑战性太大了

一点都不能错呀

但是后来我还是放弃了

你知道为什么

我研究了大量的包含复数概念的lora

也有一些大概念lora

还有尤其是最近这2个月

有一些实验性的多概念多角度lora

我自己做的别人做的我都看过

我专门   看研究

也就一直我就一直下不了决心

其实在sex from behind的时候

我就想做一个多概念多角度合一的

到底后来也没有做

你知道最大问题是什么

当然这是一个专业问题

简单的说最大的问题就是，同步拟合

就是同步拟合

因为多个概念啊，拟合没有办法

对于一个多细节的

对于一个多变化的

一个复数概念的一个lora来说

那么适合的一个拟合程度达不到，同步拟合达不到

拟合度达不到

因为你不能正面是这样子

不可能

你从正面到侧面、从握手到抓胸

这一出图就寄了

第一是拟合达不到

无论怎么样

我想来想去

因为我算个内行吧（笑）

我不炼我就知道拟合达不到（其实炼了）

同步拟合实现不了

第二个   作者的精神状态会损失一半以上

作者的精神会损失一半以上

作者的大脑会损失一半以上的san

是看不看不到的

lora就不好做了

lora就不好做了

最后我差不多半个多月的思想斗争

那是个重大决定

一个个训练集都是按照一个lora概念全包准备的

还ps、打标呢

后来说我放弃

 我自己决定放弃

一个人做的决定

一个人的思考

半个多月以后

我放弃以后

就觉得好像就没劲了

不能一个lora概念全包

就觉得好像就不想做了（哈哈哈）

开玩笑啊

就所有人都沮丧而失落（就我自己）

我也很难过

我后来还是放弃了吧

所以制作概念lora是一个学习

它对于一个追求高质量的一个创作者来说

它是破坏性的

一个lora概念全包又能怎么样

你丢掉了质量

丢掉了牺牲了出图成功率

孰重孰轻啊（点头）

一款中国宫廷服饰LOCON，服装类型有贵妃，娘娘，或者是丫鬟的宫廷装（权重越高，服装款式、级别越高，甚至有一些更像宫廷睡衣）

建议权重不要超过0.8，推荐0.6；

How to use locon 怎么使用locon：

由于webui中没有内置LyCORIS，使用前一定要记得安装插件,

https://github.com/KohakuBlueleaf/a1111-sd-webui-locon

将模型导入models\LyCORIS中即可使用

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

如果你觉得我的模型有帮助，请点亮右上角那个免费的收藏红心，不胜感激。

===========================================================

2023/6/3

V1.1Lora是一个Lora模型。

akagi_rin\(denchi-project\), star hair ornament, short hair with long locks,

akagi_rin\(bodysuit\), mechanical ears, vambrace, pauldrons,

akagi_rin\(uniform\)

稍稍凑出了点时间进行测试的一版，在衣服的还原度上仍然不好，akagi_rin\(bodysuit\)的问题在示例图里有展示了……

之后如果有机会可能会训练小鸟之翼的主角，或者是混沌之子的角色。但如果这段时间内有人做了的话我就得斟酌一下。

V1

0.55-0.6,anythingv5,LoHa.

akagi_rin\(denchi-project\),star hair ornament, twin braids

uniform

bodysuit, mechanical ears, vambrace, pauldrons, but this still requires further imorovement.

一些碎碎念 2023/5/28

作为一部所谓宅文化毁灭系的片子，电池少女在剧本上是平庸的，但并不是差劲的，尤其是跟四月的魔法少女毁灭者相比，更是展现了作为一部原创动画剧情上能够平稳落地的含金量。

然而这个片作画上的浓浓万策尽气息让它几乎不太能成为又一部小众佳作……并且让这个模型诞生的十分艰难！

我实际上一个多月前就训练了这个角色的Lora，并于近日做好了三个女主角的全部Lora，但都在服装的服从度上均表现的很微妙，我原本想打算直接发布，但经朋友提醒，这两天又用数据集赶出了这么个loha出来。

LoHa的确在角色的学习上非常优秀，比方说它甚至学到了我没扣干净的素材痕迹，以至于我不得不放弃我用恋活截图得到的一些素材。。不过在这个LoHa中你仍然能看到这一点，这显然是个过拟合的版本，但我暂时要忙于别的事无法测试，所以先发布这个模型，，

除了关键词的这些，这个角色还有护膝，创可贴，鱼形拉链等特征我没有还原，因为动画一个劲的拍面部特写和上半身导致我缺乏素材……

我去，这不是我们《圆神》里的鹿目圆吗

是3D偶像服大菲喵。推荐 prompts：

yellow eyes, white skirt, white gloves, pink hair

可根据实际需求自行调整，一般来说并不需要写全 tags。

更新日志

v3.0 数据集优化，现在生成的图片应该更高清了

v2.0 采用 LoRA 方法炼图（不是 LoHa），推荐指定背景和人物视角。如果出现大头照，建议使用 cowboy shot、full body、upper body 等 tags 规避。

推荐阅读

https://civitai.com/models/69411/guide-how-to-make-loras-with-very-few-images-advice-and-tips-english-and-spanish

训练方法

1. 在B站上找到一个3D大菲视频，尽可能地截取多个视角的大菲人物（12 ~ 20 张图即可）

2. 前往 https://www.upscale.media/g/pricing 注册一个账号，为后续批量放大图片做准备（如果你有大会员能看 4K 视频的话应该可以跳过放大这一步）

3. 前往 https://www.remove.bg/api 获取 api token，用于批量去除图片背景（这个网站每个月有50个免费次数，可以配合第 2. 步的批量放大网站）

4. 回到第 2. 步的网站，选中上传的图片，点击 Transform 创建 jobs，之后根据自己的需求选择处理方法即可（这里是 upscaler & remove bg）

5. jobs 可能需要处理大约 5 分钟（取决于你的素材量），之后把处理完的图片压缩包下载到本地即可

6. （可选）把图片压缩包上传到 https://catbox.moe/ ，之后在 colab 下载下来，这样比直接上传到 colab 速度会快很多

7. 打开 https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/main/kohya-LoRA-finetuner.ipynb ，跟着流程走就完事了

8. 由于本人是懒狗所以全程没配图，感兴趣的话可以自己摸索一下

数据集

https://files.catbox.moe/2fk46x.zip

モデルによって強度は違いますが、プロンプトにchibiを入れてディフォルメ化しないと

原作とは全く異なる長身でスリムな体型になってしまう可能性があります

The strength varies from model to checkpoints, but if the chibi is not deformed with chibi in the prompt, the body may end up with a slimmer shape that is completely different from the original.

english↓

This is lora by yamagami karuta, a member of the virtual youtuber group nijisanji.

The strength of lora is recommended to be around 0.8~1.

Trigger words:1 girl,ymgm

If you want to bring out the hat on her head, adding "tokin hat" to the prompt will work.

It also works well if you put the words that occur frequently by putting lora itself through analysis.

This may not be a lora with satisfactory accuracy since it was made as a trial.

Please enjoy it if you like.

https://www.anycolor.co.jp/guidelines/

japanese↓

これは、バーチャルyoutuberグループのにじさんじに所属している山神カルタのloraです。

loraの強度は0.8~1くらいがおすすめです。

トリガーワード:1 girl,ymgm

彼女の頭の上にある帽子を出したい場合は、「tokin hat」とプロンプトに加えると効果があります。

また、lora自体を解析にかけて頻出するワードを入れてみると効果があります。

お試しで作ったので、満足のいく精度のloraではないかもしれません。

よければお楽しみください。

LoHa of K.S. Miracle (umamusume). Please install extension, a1111-sd-webui-lycoris, on WebUI. LyCORIS, not LoCon

After installing extension, put LoHa file in your stable-diffusion-webui/models/LyCORIS folder and write LyCORIS notation in prompt to applying LoHa. <lyco:your_loha_name:weight>

All of my works is free. If you'd like to support me, feel free to buy me a coffee. 👍

NOTICE

• After careful consideration, I decided not to produce LoRA of the characters below.

  • Otonashi Etsuko, Akasaka Misato, Hosoe Junko

• If anyone wants the characters above, I will consider making them.

• Except for the above, I will make the rest of the characters in Umamusume: Pretty Derby, unless there is a special notice.

Recommended options

• LoRA weight 0.6 ~ 0.7

• Trigger words

  • Base

    • k.s.miracle \(umamusume\)

  • Original Art :

    • ear covers, blue shirt, collared shirt, blue coat, open coat, white ascot, brooch, jewelry, white gloves, half gloves, long sleeves, belt, black pants, thigh strap, tailcoat, boots

  • SSR Guts:

    • official alternate costume, ear covers, collared shirt, white shirt, dress shirt, sweater vest, yellow vest, black pants

  • Casual :

    • casual, ear covers, ear ornament, white dress, long sleeves, jeans, blue pants, brown footwear, shoes

  • Tracen summer uniform

    • tracen school uniform, summer uniform, serafuku, puffy short sleeves, purple bowtie, horseshoe ornament, sailor collar, sailor shirt, purple shirt, white skirt, pleated skirt, frilled skirt, miniskirt, zettai ryouiki, white thighhighs, brown footwear, loafers

  • Tracen winter uniform

    • tracen school uniform, ear ornament, hairclip, winter uniform, puffy long sleeves, sailor collar, purple shirt, white bowtie, horseshoe ornament, purple skirt, pleated skirt, purple thighhighs, brown footwear, loafers

Use a model derived from or mixed with animefull model.

• Settings

  • Stable Diffusion WebUI by Automatic1111

  • DPM++ Series(SDE Karras, 2M Karras, etc.)

  • about 20 steps, CFG scale 3.5~6.5

  • I recommend kl-f8-anime2.vae

  • CLIP skip = 1

  • Use Hires.fix to get higher quality image

    • Upscaler

      • Latent, Latent (nearest-exact), Latent (bicubic antialiased) or other Latent series.

    • Denoising strength 0.50~0.65

LoRA training info

• trained on sd-scripts by kohya_ss.

  • LyCORIS by KohakuBlueleaf

  • Thank you all a lot!

• based on Animefull-pruned model

• 122 images consist of

  • 4x low quality 56 images

  • 8x medium quality 40 images

  • 16x high quality 26 images

  • => 1 epoch = 960 images

• replaced character feature tags with k.s.miracle \(umamusume\)

  • 1girl, 1boy, horse ears, solo, short hair, simple background, long hair, blue eyes, horse girl, purple hair, bangs, kaito \(vocaloid\), purple eyes, long bangs, very long hair, blue hair, horns, shenhe \(genshin impact\), medium hair, light blue hair, black hair, horse tail, animal ears, low ponytail, fu hua, sidelocks

  • I think these tags are the essentials of K.S.Miracle or

  • unnecessary for training because of duplication.

• resolution : 768x768 with aspect ratio bucketing.

• other training settings is included in training data or metadata of LoRA file.

All uploaded images are generated by AOM3A1

And, special thanks to my best friend. DH.

Kotonoha Katsura, one of the protagonist of School Days

あなたの好きなチビキャラをたくさん乗せてあげてください

Put many of your favorite chibis on it!

Pretty girls in sexy clothes, in a simple style that gets along well with both 2.5d and flat-shaded models.

Tilt Shift LoHa is a LyCORIS/LoHa trained with v2.1 as a base. It can produce some pretty amazing aerial tilt shift photography on all different types of landscapes. It is not the greatest at toon-like JRPG landscapes as it's main dataset is real photographs.

LoRA trained to generate a person in public surrounded by people, such as photographers. Works with other LoRAs and various models.

UPDATE - new training and small improvements

Updated training: used LyCORIS instead of LoRA, trained on larger dataset with more realistic images, cleaned up tags.

Updated output: less likely to have a camera on character's hands, more detailed crowd, a bit more consistent, you can better increase the crowd or photographers by increasing the trigger tag's weight.

Main tags (use one or both for different effect): public, photographers

// OLD VERSION (1.0) //

Main trigger word: public Other tags used photographer,phone screen,camera,foreground

You can use your usual CFG and weight 1.0

See weight comparison 0.0 vs 1.0, minimal effect on style and composition:

Additional info

I make these models for fun and for the challenge and release them for free for everyone to enjoy, for non-commercial use (see license). The one thing I ask of you is to provide rating and feedback! That helps me understand where I can improve and helps me stay motivated to produce more content!

I accept suggestions or requests here or on Civitai discord server.

You can also check my more polished images and upcoming models on my pixiv profile.

Just an experimental piece to go along with article.
Article: https://civitai.com/articles/97
Posted only for educational and research purposes.
There are probably 20 other Rem loras out there already.

This LyCORIS has been trained after carefully selecting 59 images. All photos have been enhanced to make the most of them. I hope you like it!

Any feedback would be highly appreciated!

From my experience it runs best with Edge Of Realism, pirsusEpicRealism, epicDiffusion ,RealisticVision, CyberRealistic, Deliberate checkpoints, among others, although I think it will be just fine in other SD 1.5 models as well.

CFG scale should be between 4-7.

LyCORIS weights should be between 1-1.4

Notice: This free work is intended for art and illustration purposes of emergent technology and not intended nor' trained for making illicit images, as the author I've taken care to not use any private data in the training process or designed it for such purposes. By downloading this model you are agreeing with the aforementioned intent and that you will not use my work to create / share images of a potentially immoral and / or illegal nature.

Note #1: This is sorta a paired release with my other CaC, Bad Cop (located here https://civitai.com/models/78494). Time to see what people want and download more, that Good Cop/Nice Waifu look or that Naughty Cop/Bad Girl style ;9! I have a couple more of these Good/Bad pairings coming out eventually.

What does it do you ask? Enforcing the law and capturing criminals is the path that your waifu has decided to follow, as she is an upstanding and vigilant member of the police force. With the Locon, she becomes exactly that and will do what she can to protect the citizen of whereever-you-want.

How to use it:

(I use Sampling Method DPM++ SDE Karras and Upscaler R-ESRGAN 4x+ Anime6B).

0.7 to 1 depending on model/lora/etc combination.

Main Trigger

GCop - Is to geth the officer look rolling.

Other Useful Tags

pocket, breast pocket, shirt, pants, uniform, belt,  short sleeves, collared shirt, shirt tucked in, pantyhose, skirt, gloves, necktie, pencil skirt,  olice hat, high heels, shoes, miniskirt - All these appeared a decent amount in the dataset.

Anything Else You Want!

Extra, Extra!: Should be well trained and usable in most models. Images shown for example.

Picture Descriptions:

1: Hassaku with just this Locon.

2: AnythingV3 with just this Locon.

3: AbyssOrange with just this Locon.

4: RevAnimated with just this Locon.

5-6: Perfect World with just this Locon.

7: AnythingV3 combo'd with my released Succubus Locon.

8: AnythingV3 combo'd with my in progress Tied Till Orgasm Locon.

9: AnythingV3 combo'd with my released March 7th Lora.

10: AnythingV3 combo'd with my Chika Lora.

11: AnythingV3 combo'd with my Lucoa Lora.

How was it trained?

It was trained on animefull_final with 27 images with the usual self-pruning and tagging and a wide range of different images. I like keeping my prompts simple for the first few pictures so that you can see it without a whole bunch of mumbo jumbo. Make your pictures better than mine though!

What's next for this?

I think it's fine? LMK.

What's next that isn't this?

More character Loras. (Jill Valentine and Ada Wong from Resident Evil, Elsa and Echidna from Re:Zero, Rupee and Volume from Nikke, Spectre Tenebria from E7, Zephia from Fire Emblem Engage, Ivy from Soul Calibur 5, Heine from Combatants Will Be Dispatched, Conqueror Lilias from E7, Evelynn from LoL (base version), Tharja from Fire Emblem, Laharl-Chan from Disgaea 5, Miku from Idaten Deities, Kuribayashi Shino from GATE, Rin Sohma from Fruits Basket, Kasumi from P5, Ikumi Mito from Food Wars, Alice Nakiri from Food Wars, Lulu from FF10, Laharl-chan from Disgaea, Dominique de Sade from Vanitas, and Dos from One Punch Man.) And maybe Wilhelm van Astrea, my favorite Re:Zero husbando or a Prinny from Disgaea.

V2's of some of the previous Loras. Specifically make them more diversely acceptable with other models and LORA's.

I also have some more quite NSFW Locon's. I decided to list them as well, so if people have preferences on which to fix up first let me know. Specifically, Ruination (from League of Legends), Bowsette Transformation (Makes your waifu kinda look like Bowsette without changing other aspects), Fallen Angel, Evil Sailor Scout, Team Rocket (Make the waifu a Team Rocket Grunt), Pony Bondage (specifically with the capability of being "riden"), X-ray Vision (glasses that give you x-ray vision of a girl), Vore (Waifu mid-way being eaten specifically), Revealing Clothes (just makes normal clothing even more revealing), More Brainwashing (in ideation for this still), Slime Corruption (transforms your waifu into a slime girl), Cowify (Make your waifu into a cow-girl), Bound Yuri (two girls tied up, breast to breast and likely kissing), Tied up with vibrators (basically the waifu tied up with vibrators and being overwhelmed with pleasure), Beta Censorship (Automatically censors any images you create to make them BetaSafe), being stepped on, Robotification, Broken and Defeated (decently NSFW defeat scenario for the waifu getting abused), Petrified (I'm unsure if I should bother with this one anymore), Infection (this one is VERY nsfw, as the waifu is completely infected), Tentacles, Pillory, Latex Corruption, Combat Drone (hard to describe but basically making them a mindless sci-fi combat warrior), JAV Supervillain, Blacked, Hypnotized Harem, Hourglass Figure (just to make your waifu curvy), and Holding a Leash.

And, uh, I want to make a few SFW Locon's so I have a Pirate, Samurai, Tomboy, Delinquent (The delinquent girls in japanese schools specifically), Maid (both normal and naughty variants), forest fairy, barbarian.

Or just give me some random suggestion or priority opinions. That works too.

I've also added a couple of my extreme NSFW preview images to my twitter (https://twitter.com/TwoMoreTimes89). Let me know if you think I should work on them more for a submission here.

Note #1: This is sorta a paired release with my other CaC, Good Cop (located here https://civitai.com/models/78330). Time to see what people want and download more, that Good Cop/Nice Waifu look or that Naughty Cop/Bad Girl style ;9! I have a couple more of these Good/Bad pairings coming out eventually.

What does it do you ask? Enforcing the law and capturing criminals is the path that your waifu has decided to follow... Well, not exactly. Instead, she loves pulling over a married couple and having the husband's eyes unable to look away while she writes up a ticket for him... With her number on it. Basically, a naughty cop for a naughty waifu.

How to use it:

(I use Sampling Method DPM++ SDE Karras and Upscaler R-ESRGAN 4x+ Anime6B).

0.7 to 1 depending on model/lora/etc combination.

Main Trigger

BCop - Is to get the naughty officer look rolling.

Other Useful Tags

gloves, fingerless gloves, cleavage, police uniform, police, uniform, belt, policewoman, leotard, police hat, peaked cap, thighhighs, fishnets, skirt, fishnet thighhighs, shirt, short sleeves, pencil skirt, miniskirt, crop top, midriff, high heels, necktie between breasts,  panties, boots, earrings, jewelry, thigh boots, swimsuit, cropped jacket, highleg bikini, jacket, underboob, highleg, thigh strap, single glove, skindentation, mismatched bikini, shorts, micro shorts, long sleeves, piercing, navel piercing, cuffs - All these appeared a decent amount in the dataset.

Anything Else You Want!

Extra, Extra!: Should be well trained and usable in most models. Images shown for example.

Picture Descriptions:

1: Hassaku with just this Locon.

2: AnythingV3 with just this Locon.

3-4: Hassasku with just this Locon.

5: AbyssOrange with just this Locon.

6: RevAnimated with just this Locon.

7: AnythingV3 combo'd with my released Succubus Locon.

8: AnythingV3 combo'd with my in progress Infected Locon.

9: AnythingV3 combo'd with my released March 7th Lora.

10: AnythingV3 combo'd with my Chika Lora.

11: AnythingV3 combo'd with my Lucoa Lora.

How was it trained?

It was trained on animefull_final with 27 images with the usual self-pruning and tagging and a wide range of different images. I like keeping my prompts simple for the first few pictures so that you can see it without a whole bunch of mumbo jumbo. Make your pictures better than mine though!

What's next for this?

I think it's fine? LMK.

What's next that isn't this?

More character Loras. (Jill Valentine and Ada Wong from Resident Evil, Elsa and Echidna from Re:Zero, Rupee and Volume from Nikke, Spectre Tenebria from E7, Zephia from Fire Emblem Engage, Ivy from Soul Calibur 5, Heine from Combatants Will Be Dispatched, Conqueror Lilias from E7, Evelynn from LoL (base version), Tharja from Fire Emblem, Laharl-Chan from Disgaea 5, Miku from Idaten Deities, Kuribayashi Shino from GATE, Rin Sohma from Fruits Basket, Kasumi from P5, Ikumi Mito from Food Wars, Alice Nakiri from Food Wars, Lulu from FF10, Laharl-chan from Disgaea, Dominique de Sade from Vanitas, and Dos from One Punch Man.) And maybe Wilhelm van Astrea, my favorite Re:Zero husbando or a Prinny from Disgaea.

V2's of some of the previous Loras. Specifically make them more diversely acceptable with other models and LORA's.

I also have some more quite NSFW Locon's. I decided to list them as well, so if people have preferences on which to fix up first let me know. Specifically, Ruination (from League of Legends), Bowsette Transformation (Makes your waifu kinda look like Bowsette without changing other aspects), Fallen Angel, Evil Sailor Scout, Team Rocket (Make the waifu a Team Rocket Grunt), Pony Bondage (specifically with the capability of being "riden"), X-ray Vision (glasses that give you x-ray vision of a girl), Vore (Waifu mid-way being eaten specifically), Revealing Clothes (just makes normal clothing even more revealing), More Brainwashing (in ideation for this still), Slime Corruption (transforms your waifu into a slime girl), Cowify (Make your waifu into a cow-girl), Bound Yuri (two girls tied up, breast to breast and likely kissing), Tied up with vibrators (basically the waifu tied up with vibrators and being overwhelmed with pleasure), Beta Censorship (Automatically censors any images you create to make them BetaSafe), being stepped on, Robotification, Broken and Defeated (decently NSFW defeat scenario for the waifu getting abused), Petrified (I'm unsure if I should bother with this one anymore), Infection (this one is VERY nsfw, as the waifu is completely infected), Tentacles, Pillory, Latex Corruption, Combat Drone (hard to describe but basically making them a mindless sci-fi combat warrior), JAV Supervillain, Blacked, Hypnotized Harem, Hourglass Figure (just to make your waifu curvy), and Holding a Leash.

And, uh, I want to make a few SFW Locon's so I have a Pirate, Samurai, Tomboy, Delinquent (The delinquent girls in japanese schools specifically), Maid (both normal and naughty variants), forest fairy, barbarian.

Or just give me some random suggestion or priority opinions. That works too.

I've also added a couple of my extreme NSFW preview images to my twitter (https://twitter.com/TwoMoreTimes89). Let me know if you think I should work on them more for a submission here.

LyCORIS of Indiana A, prolific modeler for MetArt and others from 2008-2016. Trained at 768x768 with 112 images, a mix of close-ups, waist up and full body shots, wearing a variety of clothing (or not ;) ).

Use indiana_a as the trigger, can use freckles also to control how much they are visible.

Sample images were generated with realisticVisionV20, with 0.65 strength then inpainted at 1.0 strength.

x光 midjourney风格lora，权重设置在0.5-1.3之间，底模选取realisticvision

关键词：xguang，illustration，

可以产出，人物，风景，动物，汽车等等

LoHa of Satono Crown (umamusume). Please install extension, a1111-sd-webui-lycoris, on WebUI. LyCORIS, not LoCon

After installing extension, put LoHa file in your stable-diffusion-webui/models/LyCORIS folder and write LyCORIS notation in prompt to applying LoHa. <lyco:your_loha_name:weight>

All of my works is free. If you'd like to support me, feel free to buy me a coffee. 👍

NOTICE

• After careful consideration, I decided not to produce LoRA of the characters below.

  • Otonashi Etsuko, Akasaka Misato, Hosoe Junko

• If anyone wants the characters above, I will consider making them.

• Except for the above, I will make the rest of the characters in Umamusume: Pretty Derby, unless there is a special notice.

Recommended options

• LoRA weight around 0.7

• Trigger words

  • Base

    • satono crown \(umamusume\)

  • Original Art :

    • ear ornament, black gloves, elbow gloves, collared shirt, suspenders, bare shoulders, yellow ascot, green shirt, sleeveless shirt, black shorts, highhighs, thigh boots, black footwear

  • Tracen summer uniform

    • tracen school uniform, summer uniform, serafuku, puffy short sleeves, purple bowtie, horseshoe ornament, sailor collar, sailor shirt, purple shirt, white skirt, pleated skirt, frilled skirt, miniskirt, zettai ryouiki, white thighhighs, brown footwear, loafers

  • Tracen winter uniform

    • tracen school uniform, ear ornament, hairclip, winter uniform, puffy long sleeves, sailor collar, purple shirt, white bowtie, horseshoe ornament, purple skirt, pleated skirt, purple thighhighs, brown footwear, loafers

Use a model derived from or mixed with animefull model.

• Settings

  • Stable Diffusion WebUI by Automatic1111

  • DPM++ Series(SDE Karras, 2M Karras, etc.)

  • about 20 steps, CFG scale 3.5~6.5

  • I recommend kl-f8-anime2.vae

  • CLIP skip = 1

  • Use Hires.fix to get higher quality image

    • Upscaler

      • Latent, Latent (nearest-exact), Latent (bicubic antialiased) or other Latent series.

    • Denoising strength 0.50~0.65

LoRA training info

• trained on sd-scripts by kohya_ss.

  • LyCORIS by KohakuBlueleaf

  • Thank you all a lot!

• based on Animefull-pruned model

• 41 images consist of

  • 4x low quality 17 images

  • 8x medium quality 12 images

  • 16x high quality 12 images

  • => 1 epoch = 356 images

• replaced character feature tags with satono crown \(umamusume\)

  • 1girl, solo, black hair, simple background, short hair, horse girl, crown, hair between eyes, horse ears, grey eyes, very long hair, horse tail, grey hair, multicolored hair, animal ears, white hair, twintails, long hair, mini crown, bangs, asymmetrical bangs, green eyes, purple hair, blue eyes, side ponytail

  • I think these tags are the essentials of Satono Crown or

  • unnecessary for training because of duplication.

• resolution : 768x768 with aspect ratio bucketing.

• other training settings is included in training data or metadata of LoRA file.

All uploaded images are generated by AOM3A1

And, special thanks to my best friend. DH.