大多数模型适用性良好，权重0.4-1都可以，根据权重可以从漏洞装进化到战败。

Most models have good applicability and can be used with weights ranging from 0.4 to 1. Depending on the weight, it can range from a leaky outfit to a defeat.

在chilloutmix模型上实现2.5d人物

Here comes our favorite Yui (2015 version) before her intensive surgeries in the last few years.

For how I train LORAs, see description of JAV Anzai_RI○N | Stable Diffusion LORA | Civitai

Must turn on Hires.fix

Hires steps:17, denoising:0.5+

Use DPM++ 2M Karras or DPM++ SDE Karras

Strength: 0.6-0.7

She might look like someone from real life, but she is not what you think. She is a totally fictional character.

Checkpoint used for samples: Basil Mix and KanPiroMix

VAE: vae-ft-mse-840000 from StabilityAI

Legal & Risk

It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.

本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时，请务必遵守当地法律法规，勿侵犯他人的名誉权、隐私权或肖像权。

When using this LoRA

1. start the weight at 0.2. If you don't have many prompts, it'll work well enough. Hyper penis and futanari must be included as triggers.

2. If it doesn't work properly, highlight the trigger word first and then check for changes. If you want a bigger penis, try adding big penis, long penis, and super penis to the prompt. I recommend that the LoRA weight not exceed 0.5. It affects the body of the picture and is likely to move away from the picture you want.

p.s.) There may be awkward parts or parts that you don't understand because you are using a translator. If you ask me a question, I'll answer it hard.

Hi, guys, I'm back again! As you can see, this is a model depicting the delivery of pregnant womans. Its effect looks very good. Actually, it has an early version, but sadly, it only survived for 5 hours in Civitai.

Now I have conducted intensive training on it to support the following scenarios.

• Delivery of head(Similar to shussan kai,But it can control whether to draw amniotic fluid or not)

• Delivery of body(It still needs optimization, it only looks a bit more than the head)

• Delivery of whole body

• Draw umbilical cord only(tag:umcord)

How to use?

Please draw a regular bare pussy female image first, and then send it to the part redraw function of img2img. Use this Lora model to redraw the pussy area.

you must addon tag:

1girl,birth

then you can choose:

headbirth,bodybirth,finalbirth,amniotic fluid,umcord

Why cannot use in txt2img?

Because some tags are not cleared, Drawing pictures directly may affect the style of painting.

Have u a good fun!

Do you want to support me? Consider being a patron on Patreon to vote on train polls and access stuff early. ❤️

Trained on 100 image of new SF6 outfit and 768x768 resolution. For better quality, use hires fix to generate images.

Trigger Word is ChunliMS

Tags you'll need for her looks are

hair bun, double bun, eyeshadow, makeup, chinese clothes, pelvic curtain, solo, dress, bracelet, ribbon, pants, blue pants, makeup, short sleeves, wide hips, thick thighs

Use the LoRA weight around 0.8 and 1 for the best results

Discord server feel free to drop and say hi ^^

I gathered representative Korean idols and realized a fictional character.

<lora:koreanBoys_v10:0.8>

Trigger word is '1boy, solo, shirt, necklace' etc.

I hope you enjoy it.

秋月型防空駆逐艦その八番鑑、冬月だ。戦いの後半で就役して、姉妹艦と第４１駆逐隊を編成。大型艦や艦隊護衛を務めだな。最後の出撃は、第２水雷戦隊の一翼として、涼月とともに大和の護衛についた。そして今も、僚艦涼月と北九州の防波堤の一部として、みんなを守っている。

German supermodel and actress. Fashion designer Valentino once commented on her physical resemblance to Marlene Dietrich.

A beautiful ruassia model,you deserve it!!

你可以在这个网站赞助我，你的赞助就是我创作的激励

You can sponsor me on this website, and your sponsorship is my motivation for creating

https://afdian.net/a/NovitiateAlchemist

This version has better sex content

这个版本有更好的sex内容

• Full Body does not work well, please use Image to Image to increase the resolution to solve it！！！！

• Updated 2 outfits:yunxi dress,nude(Please add when using yunxi nude)

• 更新了两套服装:yunxi chinese dress(云溪中式服饰),nude(使用时请加上yunxi dress)

• tag:yunxi dress,nude

• yunxi v2版本经过测试已经完好,只是分辨率过低全身的脸部效果不佳，请使用图生图功能提高分辨率即可。

• YUNXI V10版本效果更佳,支持更多模型

• 在一些模型使用不佳时，请你适当降低yunxi lora 的权重

• 在使用其它的lora模型时，如果效果不佳，请适当降低权重

• 推荐底模使用perfect world，tag使用photorealistic,realistic,photo referenced,best quality, ultra high res,(detailed face and eyes),等提示词（这些提示词可以不添加，但效果可能会不好）

• 为了更还原角色你可以添加更多原著形象细节

• 权重建议如下图

This "LoRA," a character trained based on the game "Tower of Fantasy" Everyone is welcome to use it and give me feedback!

More AI artwork here
https://twitter.com/AISexy_18m

基于chilloutmix，推荐0.6-1.0，采样步数20-30, DPM++SDK Karras。

搭配其他脸型lora使用效果更佳。

Asuka, she is so cute:)

And it also fits well with lora:moxin

有人请我训练这个角色的lora，我不认识这个角色

但她足够漂亮，值得一个lora

练的比较硬，还原度足够

a lora to turn your image to a scene like lego

Kohya GUI, 400 Images, 15 steps each.

yami is a 3d character i created by generating images and then trained the model to draw her from the generated images.

I advice to put this model weight always to 1 for best results.

Look at example images if you are looking for good prompts.

Comment and post your best images, i'm waiting for them!!

彼亚塔曼个人lora，作者boxcat。推荐模型expmixLine_v20或mixpro系列，权重0.55-0.8，提示词byjg。请勿商用。

如果想像一点可以在0.7左右调试，禁止商用

trigger word : j3nn1f3r

recommended strength : 0.85 to 0.9 -ish

you can add "lipstick" to increase likeness

going below 0.75 lora strength will break the likeness

thank you for visiting my post <3

Training LoRas is expensive, if you like my stuff pls consider supporting me on Kofi

It generates siege engines with all the AI craziness included.

I offer all my content at no cost to you. Should you wish to show your appreciation and support, kindly consider buying me coffee via the following link: https://www.buymeacoffee.com/konyconiN. Your generosity is greatly appreciated!

Golfer.

Legal & Risk

It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.

In principle, I do not support the use of this model to generate any adult content. When using this model, users must abide by laws and regulations, and do not infringe other people's reputation, privacy or portrait rights.

Best outcome with ChilloutMix and variants

• 512,768 image

• Use Euler, DPM++ 2M Karras, DPM++ 2S a Karras or DPM++ SDE Karras

• Use 20-35 sampling steps

• Weight: 0.7-1.2

• CFG: 5-7.5

  If using Hires:

• Hires step: 20

• R-ESRGAN 4x+ or Latent

• Denoising: 0.5-0.7

After download this model, try to share with us about your findings.

Comments and suggestions are needed for the development of this model.

Again, not for commercial use and she is not an existing person.

Could NOT get the heterochromia to work most of the time. Adding it to captions makes her eyes overly blue. Tried specifying the left and right. Tried triple-quadruple bracketing/weighting and changing token the order. Nothing worked.

And the rare time things kind of worked, it gave her too much cleavage to post on here.

Maybe adding it to the training captions? Or more close up images? Or more training? Or I will just blame the model checkpoints.

Esthetic 90s style (and a little 80s)

New version with increased amount of data 1.3RW, 1.4RW, 1.5RW, 1.6R on Patreon

A small version comparison:

when I release version 1.3, versions 1.2 and 1.2R will be released to the public

Hands finally reached the style of the 90s. There were many difficulties, dozens of versions, but I found a golden mean, from which I will continue to build on future updates.

just use hires fix as much as your video card is capable of, so as not to give an error, with 8gb of video memory it is 512x768 with hires fix 2, upscaler Latent (nearest-exact), the picture will be much more detailed. If you want a style close to the original but with more detail, don't use Denoising strength above 0.55, but to avoid artifacts, do not use too low a value, optimally 0.5

LORA weight 0.6

CFG Scale - 7

Eta noise seed delta 31337

Use Abyssorangemix3aom3_aom3a3.safetensors With these negative prompts:
(worst quality, low quality, extra digits:1.4)

Trigger Words: First, depending on what you need, trigger words are not required at all, read:

1990s \(style\) - most often makes enlarged eyes and detailed pupils, as in anime Slayers or Sailor Moon for example.

1980s \(style\) - there are some 80s arts, when used together with 1990s \(style\) sometimes very interesting results are obtained

retro artstyle - additional optional trigger

Combined with my other LORA

How to use LoRA's in auto1111:

• Update webui (use git pull like here or redownload it)

• Copy the file to stable-diffusion-webui/models/lora

• Select your LoRA like in this video

• Make sure to change the weight (by default it's :1 which is usually too high)
  
  *Information taken from Lykon
  
  Please post your work with or without comments, it will help me improve. Thanks!
  
  If you like my work, click on the heart above, I will be pleased :3

2011 Racing Miku Clothes

Recommended weight 0.7~0.9

Please use the following prompt words:bare shoulders, see-through,thigh_boots,elbow_gloves,race_queen,white legwear,

0.8-1左右合适

一个实验性质的lora，我将14张风格截然不同的图片放在了一起进行了训练

（其中有场景的，有物件的，还有人物的）

由此得到的lora总能出现很有意思的结果

以后我都会将以这个方式得到的自己满意的lora发布在这个栏目中，

点个♥不迷路

玩的开心

For an experimental lora, I put together 14 images of very different styles for training

(some scenes, some objects, or some people)

The resulting lora always produced interesting results

I'll be posting the lora I get in this way in the future.

Click ♥ and don't get lost!

Have fun!

An attempt to modernize old photos, but I don't feel it got enough of her likeness.

Maybe it's the fault of the model or the prompt?

soultide-Asuna 星见亚砂

Currently includes Bridgette, Courtney, Gwen and Lindsay from Total Drama.

I will probably add more characters in the future, as well as train the model for more epochs.

I got some good results using anythingv4.5, starting at low resolution and upscaling in img2img with denoising strength around 0.6.

soultide-Mist炼的不是很好

13 Sentinels Aegis Rim

Natsuno Minami

十三机兵防卫圈

南奈津乃 （みなみ なつの）

看了C站没有南的LORA，却有其他三个女角色，随便炼炼

仅供娱乐，请勿用于商业用途

bilibili 赤星ASUKA

This is a LoRa about tracen school uniform.There are two versions here, one for summer uniform and the other for winter uniform. Download according to your needs.

Suggest weight:around 1

The trigger words are "summer uniform" and "winter uniform",

optional:thighhighs.

Use the correct prompt word according to the version you downloaded.

Use highres fix. to get better results.

If unwanted parts appear in the generated images, please add them in the negative prompts, such as "animals ears."

In theory, it can be used together with other LoRa models (as shown in the sample images), But the effectiveness may be limited by the way LoRa was trained.

So,I suggest adjusting the prompt and LoRa weights, and recommend using sd-webui-lora-block-weight to set the weights.By the way, I have already adjusted the weights of this LoRa to what I think is suitable using SuperMerger.

Since the dataset consists entirely of anime images, and the model is based on the novel-anime-full, I cannot guarantee that it will work well on other realistic models, such as chilloutmix.

Only trained on 19 images for 330 steps (<9 minutes), but turned out pretty decent.

I still don't understand what makes one LoRA work out and another not?

Good day commander! This time I have tried different training methods, adjusted the training set and optimizer to get better style and details, there will be some overfitted, but the result is very satisfactory, this time you can trigger the same posture and expression as the characters in the game through tag, such as adding honolulu (azurlane), tsundere, you can get girls similar to Honolulu, recommended weight 0.3~0.6,hires fix is also necessary for getting better result, enjoy!

here is all the game character you can trigger:

taiho\(azur lane\)

stlouis\(azur lane\)

Honolulu\(azur lane\)

Pola\(azur lane\)

zara\(azur lane\)

Surcouf\(azur lane\)

veneto\(azur lane\)

agir\(azur lane\)

这次尝试了不同的训练方式，调整了训练集以及优化器，以获得更好的风格以及细节，会有一些过拟，但结果令人满意，这次可以通过tag 触发和游戏中人物同样的姿态和表情，比如加honolulu\(azurlane\) ,tsundere,就可以获得和火奴鲁鲁相似的女孩，推荐权重0.3~0.6，建议开高清修复以获得更好的结果

Too difficult to shrink, use with caution

It's a bit hard to change clothes color

Form for requests.

I also sometimes do polls on my Patreon for the next LORA (no payment required).

A weight of ~0.7 is recommended.
Feel free to experiment with other weights and share your results if you want to!

The main prompt is "Sophia"
If you are having issues with the hair color, add "blue hair" to prompts.

The dataset for this one was again quite limited, so her outfits won't be replicated 100%.

For her outfits try:

Sophia (Pious Pilgrim)

Sophia, twintails, gloves, hat, SophiaDress

Sophia (Enlightened Priestess)

Sophia, twintails, white dress, white gloves, elbow gloves, hair flower, circlet

Sophia (Pilgrim on a Short Break)
Sophia, long hair, beret, coat, open coat, sweater dress, belt

Let me know if you have any issues.

Used embeddings for sample pictures:

EasyNegative

bad_prompt_version2

If you have any feedback or questions just write in the comments or hit me up on reddit or discord.

GenshinImpact Nilou LoRA

The reproduction is incomplete, but it is fairly usable.

Please be aware that the hair often fuses with the transparent costume.

If you don't include a quality tag, it is often juvenile.

If it doesn't work, the only way to get it to work is to do it by doing it a number of times.

And don't get hung up on sample prompts.

Thank you for your patience.

原神インパクト ニィロウ LoRA

再現度は不完全ですがまあまあ使える出来にはなってます。

透明な衣装と髪の毛が融合することが多いのでご承知おきを。

クオリティタグを入れないと幼くなることが多いです。

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

SamplePrompt

Nilou GenshinImpact, <lora:nilou:0.4> 1girl ,kawaii, aqua eyes eyeliner, orangered hair shiny hair, horns fake horns, long hair, low twintails, parted bangs sidelocks, medium breasts navel skindantation, see-through veil, circlet blue gemstone, neck ring, white headwear, brooch, bracer, blue nails, nail polish, thighs, gladiator sandals, thighlet :D blush midriff puffy long sleeves detached sleeves, lightblue harem outfit blue skirt bare shoulders, gold trim

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 1168806266, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

LOGH an old school japan scifi anime 1988

https://en.wikipedia.org/wiki/Legend_of_the_Galactic_Heroes

like the style of its space ship desigh

https://gineipaedia.com/wiki/Category:Alliance_vessels

weight 0.3~0.7

cfg <5

and you ca test it

LoRA based on Defaultz(_17) art style

• Trained on YesMix 1.6.

• I recommend weights between 0.6-0.9. I liked 0.7 most in my tests.

• I also tested the LoRA on a few other ckpts like CameliaMix and LibMix, and while the LoRA doesn't transfer the style to those, the outputs are still quite nice imo. So I recommended sticking with YesMix or ExpMix.

• Since most of training images are in like a magazine cover style, you might want to drop those tags into the negative.

• I recommend using keywords like various part tattoos, jewelry, piercings.

• While this is mainly a NSFW LoRA, it still can do normal stuff pretty well.

use to 1 ~ 0.8

Since it was trained with very few images, overfitting will occur.

Different hair colors appear frequently. Use the prompt 'white hair'.

However, you can try different outfits by writing prompts.

I forgot and didn't train bunny costume, will add it soon.

1.Normal

ephnel, leotard,

or

Negative

jacket

Can generate some images like product blueprints.

Common words can be: blue print,Disassembly diagram.

MOdel: anything 4.5 pruned.

As the first version does not look very good, I will announce the new version as soon as possible.

No.1xx 是一些我完全不认识的人，个别知道的我或许也因为懒得改名就放在这个系列里。

No.2xx 大概会是我个人喜欢的明星或者博主。

lora强度推荐0.6-0.7。chilloutmix出图，词条看示意图，负面我用了EasyNegative.pt文件。

素材来自网络，素材主要为侧脸。

No.135大概是被管理员删了。

公共频道：t.me/dawn66666666

GenshinImpact Hu-Tao LoRA

The reproduction is incomplete, but it is still usable.

The head pattern is inaccurate, but the atmosphere is there, so patience is important.

If it doesn't work, the only way to get it to work is to do it by doing it a number of times.

And don't get hung up on sample prompts.

Thank you for your patience.

原神インパクト 胡桃 LoRA

再現度は不完全ですがまあまあ使える出来にはなってます。

頭の模様は不正確ですが雰囲気は出ていますので我慢が大切。

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

SamplePrompt

Hu_Tao GenshinImpact, <lora:hutao:0.4> 1girl, kawaii, grin twintails, bangs, very long hair, multicolored hair, hat, flower, black headwear, hat flower, hair between eyes, bracelet, plum blossoms, brown hair shiny hair, red eyes, symbol-shaped pupils, thighs, ring, jewelry grin, smile, blush black shorts, short shorts, black footwear, white socks, chinese clothes, shirt, long sleeves, wide sleeves, tailcoat

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 3903726150, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

Lora Model of Alexandra Daddario images used were from her instagram

Images Created selecting none or nearest in upscaler and just upscale by 2 Weight used 1.0 or less works well

Images created using VisiongenRealism older version

1350 steps of training may do a version 2 with over 2000 steps

Check Out my work here!

RDigitalArtist - Hobbyist, Digital Artist | DeviantArt

And here

RDigitalArtist | Creating Digital Art of hot babes and Models for SD | Patreon

触发词：yaqian

推荐权重为0.3 ~ 1

采样器：DPM++ SDE Karras(推荐)、DPM++ 2M Karras

基础模型推荐使用 ChilloutMix

• Give your images better light, better shading

• Use descriptions such as "volume light", "shadow", etc to enhance this effect

• 为你的画面增加更好的光线和阴影

• 使用“体积光”、“阴影”等描述来增加这种效果

• 欢迎加入AI绘画模型博物馆QQ群

• 1群：791222249

• 2群：599962550

11th Release, 堕姫 - Daki from Kimetsu no Yaiba.

Put everything in Filter for seeing all images!

Examples Image is now active, Click on Show more for Weight & Trigger recommendation, Also click on About This Version at Right Corner for more Info.

I didn't not find a Lora on Daki & Not so much Art on it, so again i created a LORA Like MTSpider.

This Lora is trained on 114+ Images with Good quality, I sorted everything manually, Only picked the best possible. (SFW + NFSW)

AnimeFullFinalPrunedFP16 used, so he work with the majority of Anime Model, you can see in examples images. (Abyss Model + 7th Corneo's Heaven's Model.)

Well, he can do Default Skin with Black hair & Demon Skin with Multicolored Hair White & Green Hair.

About Trigger Words & Prompts: DakiV4 (For Trigger) & for prompts just refer to examples image.

Weight recommended: 0.6-1. - Sometime weight at 1 is too strong so the Best result is 0.8.

Negative Embedding used: EasyNegative, BadPrompt2.

Leave a Review & Share you're image in comments if you appreciate this Work :).

Trush V1.0

[模型细节调节LoRa︱模型手术刀]

说明 | specification

• This model uses a special training method

  采用特殊的训练方法

• This lora can be used to check the garbage model and adjust the details of the model by appropriate weight arrangement.

  本lora即可用于检验垃圾模型，适当安排权重也可以调整模型细节。

• According to their own model to use positive and negative weight for adjustment, the actual test results are good

  根据自己的模型使用正负权重用于调节，实际测试下来效果不错

• The problem is, people who know how to use this lora don't need to read the manual, people who don't know how to use this lora read it and won't use it, so I don't need to spell it out here, right

  问题是，会用这个lora的人不需要看说明，不会用这个lora的人看了说明也不会用，所以我没必要写明使用方法

• Whether to choose A or B needs to be seen according to the specific model, some use A better effect, some use B better effect

  选取A还是B需要根据具体的模型来看，有的用A效果更好，有的用B效果更好

使用 | Use:

• Adjust the model by selecting positive and negative weights and values according to your own model situation

  根据自己模型情况选择正负权重以及数值，即可完成对模型的调整

• This lora can be used with any prompt word for any model.For example, you can use it with Rainbow lines style LoRa to correct details

  此lora对于任何模型任何提示词都可以使用，例如你可以和彩线法LoRa一起使用达到如图所示的效果

  

• Note that this model is careful to turn on too high a weight, otherwise the output graph will collapse completely

  注意，本模型谨慎开启过高的权重，否则会产生过多的细节造成输出图完全崩坏的效果（权重过高），或者使画面丢失细节（负数权重过低）

  

• May not be compatible with partial LorAs, such as txmfkr

  可能不兼容部分lora，例如txmfkr等

局限性 | Limitations

• Not compatible with live action models (not tested, training set has no map)

  不兼容真人模型（没测过，训练集也没对应的图）

• There may be failures or crashes that are not covered by the training set

  可能有训练集未涉及的存在导致失效或者崩溃

• As for the merged model with excessive partial fitting, it will collapse once used and cannot be used at all

  对于部分拟合过度的乱融的模型一用就崩，根本无法使用

• The upper limit is too low(Use some of the more useful models to know, generally is not touched)

  上限过低（用过一些比较好用的模型就知道了，一般是碰不到的）

• This model is not suitable for fixed proportion ckpt, only suitable for external adjustment

  本模型不适合固定比例容入大模型，只适合外置调节使用

特别鸣谢 | Special Thanks

• GitHub - bmaltais/kohya_ss：训练UI，非超好用！

• Automatic1111: SD-WebUI，帮助处理训练集，当然很大一部分图是我自己用PS裁的

• Danbooru：打标器，帮助标注训练集图片，当然很大一部分图是我自己标的

• 秋叶的甜品店（频道）：让我发现了这个LoRa的训练方法，以及促成了这个LoRa的制作

祝大家玩得开心＜3

• 你可以随意将本模型合并到其他地方，但如果你共享该合并模型，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.

Lora model customization: click here

Animation series: Click here

Version Description

V100: Can change clothes, hair color, eye color, support various main models, support 512-1280 resolution

The model is made based on 国风3 GuoFeng3 | Stable Diffusion Checkpoint | Civitai modal。

The image source and training code come from the Internet, and the model is only used for exchange of scientific research interests. If there is any infringement, please contact to delete, thank you.

图源和训练代码来源于网络，模型仅作为科研兴趣交流使用，如有侵权请联系删除，谢谢！

画像ソースとトレーニング コードはインターネットから取得され、モデルは科学研究の利益の交換にのみ使用されます.侵害がある場合は、連絡して削除してください。

I used Naibasil as a learning model to learn 20 illustrations of Rock dove by Kemono Friends, includingofficial illustrations.

When using a Lora, if you set the minimum setting of 0.6, most of them come out well.

Below is a prompt for Rock dove and basic clothes. It's better if you add Rock_pigeon_(kemono_friends).

• basic clothes

  rock_pigeon_(kemono_friends),bird_girl,head wings,long hair, multicolored hair, grey hair, black hair, bird tail, bird wings, orange eyes,brown eyes,hair between eyes,large breasts,heart
  
  multicolored clothes,shirt,vest,pleated skirt,pantyhose, scarf, shoes, short sleeves, wristband,

GenshinImpact qiqi LoRA

The reproduction is incomplete, but it is fairly usable.

The problem is that the hats and the fudges tend to fly off somewhere and multiple braids are often produced.

If it doesn't work, the only way to get it to work is to do it by doing it a few times.

And don't get hung up on sample prompts.

Thank you for your patience.

原神インパクト 七七 LoRA

再現度は不完全ですがまあまあ使える出来にはなってます。

帽子とお札がどこかへ飛んでいきがちなのと三つ編みが複数本出ることが多いのが難点です。

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

SamplePrompt

qiqi GenshinImpact, <lora:qiqi:0.5> 1girl, kawaii, blush :o ofuda qing_guanmao jiang shi long_sleeves, necklace, purple_eyes, jewelry, dress, bandaged_leg thighhighs, bangs, lightpurple_hair shiny hair, braided_ponytail, hair_ornament, white_legwear

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 493943350, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

We will comply with the with the secondary creation guidelines.
https://vspo.jp/guide/
Please read the secondary creation guidelines.

I used Naibasil as a learning model to learn 45 illustrations of Black-tailed gull by Kemono Friends, includingofficial illustrations.

When using a Lora, if you set the minimum setting of 0.6, most of them come out well.

Below is a prompt for Black-tailed gull and basic clothes. It's better if you add Black-tailed _gull_(kemono_friends).

• basic clothes

  black-tailed_gull_(kemono_friends),head wings,multicolored hair,white hair, red hair, black hair, blonde hair, yellow eyes, hair between eyes, short hair,
  
  yellow pantyhose,white dress, sailor dress,pink bow, pink bowtie,sailor collar, short sleeves,frills,yellow shoes

GenshinImpact Klee Lora

The reproduction is incomplete, but it's a fairly usable product.

If it gets taller, add a word indicating toddler to the prompt.

If it doesn't work, the target method is to just do the numbers.

And don't stick to the sample prompts.

Thank you for your patience.

原神インパクト クレー LoRA

再現度は不完全ですがまあまあ使える出来にはなってます。

背が高くなってしまう場合はプロンプトに幼児を示す単語を加えてください。

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

SamplePrompt

Klee GenshinImpact, <lora:klee:0.4> 1girl, kawaii petit, blond hair shiny hair, low twintail, red eyes, skindantation smile blush cabbie hat, white shorts, long sleeves, pointy ears, red dress, backpack, red headwear, brown gloves

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 2648303014, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

fightver：

该版训练集本使用了80%新的图片，使居合的姿势更加具有动感，一定意义上可以算是另一个lora了。需要站着的pose建议使用V1，喜欢战斗姿态则选择新版本。新版本对真人的适用性有所降低，推荐二次元模型。

This version of the training set uses 80% new images, making the pose more dynamic, and in a certain sense, it can be considered as another LORA. For standing poses, it is recommended to use V1, while the new version is more suitable for fightver. The new version's applicability to real people has decreased, so it is recommended for use with 2D models.

/////

鉴于居合难以使用prompt调用，于是做了一个居合的lora。

trigger words 是battōjutsu, (holding weapen:1.3)，你也可以加上 sheath, unsheathing,

权重0.5-0.9，三次元二次元模型都行。

如果你使用Hires. Fix的话，Denoising需要调低到0.4以下。

 In view of the difficulty of using prompt for battōjutsu, a lora for it was created. The trigger words are battōjutsu , (holding weapen:1.3), and you can also add sheath , unsheathing. Weight can be set between 0.5 and 0.9, and both 3D and 2D models are suitable.

If you are using Hires. Fix, denoising may need to be set below 0.4.

我不知道为什么加不上图片prompt，原图prompt参考：

1.(masterpiece:1.5),(ultra-detailed), (high quality:1.3), ((dramatic lighting)),sharp focus,battōjutsu, (holding weapen:1.3), 1girl, solo, black hair, sheath, unsheathing,(white:1.2) kimono,open kimono,big breast,abs, (ulzzang-6500-v1.1:0.6),(katana:1.2), (Long Arms:1.2),Cherry Blossoms Falling,big moon,moon light,reeds, (pureerosface_v1:0.5),high ponytail,

2.best quality, (masterpiece:1.5),(ultra-detailed), (high quality:1.3), (high resolution),best quality ((masterpiece,best quality, detailed)), ultra-detailed, detailed hair, hard rim lighting,((dramatic lighting)),sharp focus,battōjutsu, (holding weapen:1.3), 1boy, male focus, solo, old, old man, beard, facial hair, unsheathing, sheath, japanese clothes, manly, realistic, open clothes, samurai, white hair,Cherry Blossoms Falling

You can get an idea about the specifics of this LoRA by looking at example images.

The base model is Suzumehachi.

---Kallen V2.0 4月8日(April 8th)---

最好的卡莲Lora!(The Best Kallen Lora Ever!)

Lora weights: 0.5-0.7(0.65感觉最好,0.65 maybe the best)

触发词(Trigger Words)

公主裙(Formaldress):kallencodegeass, formaldress

兔女郎(Playboy bunny):kallencodegeass, playboy bunny

驾驶服(pilotsuitr2):kallencodegeass,pilotsuitr2

校服(schooluniform):kallencodegeass,schooluniform

工作服(workclothes):kallencodegeass,workclothes

黑色骑士团服(blackknightuniform):kallencodegeass,blackknightuniform

如果有头戴,建议加上red headband来加强.

(Suggest to add promt:red headband,if you need Kallen to wear headband)

---Kallen V1.0 3月29日(Match 29th)---

生日快乐卡莲!!!

Happy Birthday Kallen!!!!

动画Anime: Code Geass 反叛的鲁路修/ CODE GEASS Lelouch of the Rebellion/コードギアス 反逆のルルーシュ

生日Birthday: 3月29日/March 29th

本命Name: 红月卡莲/紅月カレン/Kouzuki Kallen

全名Full Name: 卡莲·休妲菲尔特/Kallen Stadtfeld/カレン·シュタットフェルト

Lora weights: 0.5-0.7

触发词(Trigger Words)

公主裙(Formaldress):kallen stadtfeld, formaldress

兔女郎(Playboy bunny):kallen stadtfeld, playboy bunny

驾驶服(pilotsuitr2):kallen stadtfeld,pilotsuitr2

效果最好的公主裙，其次兔女郎，最差的是驾驶服。

花了很多时间准备训练用的素材，在做lora的过程中，突然发现卡莲的生日快到了，所以加班加点在卡莲生日之前赶出来的第一版。

第二个版本会加入更多素材提高lora质量和训练更多衣服（校服，黑色骑士团服等）。

The best effect of this lora is the formaldress one , playboy bunny is the second , and the worst is pilotsuitr2.

It took a lot of time to prepare the raw material. But I found that Kallen's birthday is coming up few days ago. So I rushed out the first version of lora before Kallen's birthday.

After the first version of lora, I will train more clothes of Kallen (such as : school uniform, Black Knights,etc) and more materials will be added to improve the quality of lora.

recommended weight: 0.7-0.9

trigger words: long pony tail, up pony tail, white cape, black short pants

This model is trained on public illustrations of Bison(Arknights).

It works well with non-style anime models such as Anything Family, for pastel-style models and others (i.e., SoftPastel, Yaoi and Pastel2D) you will get a detailed kemono cow-boy but not the same as Bison the character.

Baseline is that the stylized model can draw cow ears but cannot draw cow horns unless forced it to, so you could use this on other models to draw yourself a cow boy.

It's my first Lora trying so if you have any suggestions please tell me. With much appreciation.

What's this LoRA

This LoRA is trained on NSFW images!

This is a LoRA trained on a particular style. It creates adult illustrations of cartoon character with crazy proportions and an amateurish vibe.

The images in the examples are created using only this LoRA and different models (mainly my custom:

• https://civitai.com/models/24149

• https://civitai.com/models/26332

• https://civitai.com/models/28322

• https://civitai.com/models/32022

You don't need to specify a word to trigger this LoRA.

The LoRA also works on NSFW.

The images are created using different weights (you can check on the image generation panel)

Flavors

There are 2 different flavors available (both trained starting from Mistoon_Anime):

• Style: this is trained on AI generated pictures with an amateurish adult style. At high weight, it outputs an image with exaggerated proportions.

• Huge: this is trained on cherry-picked pictures generated using the Style flavor. This has a more consistent style biased towards huge breasts.

• Firm: this is trained on cherry-picked pictures generated using the Style flavor. This has a more consistent style biased towards smaller but firmer breasts.

Style:

Huge:

Guides

Learn how to create pictures like mine with my step-by-step tutorials:

• Beginner's Guide

• Prompting

• My Workflow

• Inpainting

If you want to create images with a high LoRA weight, I highly suggest you to check the tutorial.

Support

If you want to support my work, you can check out my music here:

https://inzaniak.bandcamp.com/

You can also check my AI Art accounts, where you'll find daily pictures made with my models (and previews of new ones coming):

• Instagram

• DeviantArt

Demon Lord, Retry! - Girlpack LoRA

Use Negative Prompts! Try lower LoRA Strengh

All Preview Images are 0.4 Strenght!

Trained for max Flexible Prompting :)

Version 0.1 Girls:

1. aku

2. angel

3. killerqueen

4. kyon

5. luna

6. mikan

7. momo

8. tron

9. yukikaze

10. yuu

(9. yukikaze is a trap ;D)

Trained on: W:480 H:720

Triggerwords:

"name" - see list

anime art style - Trigger for Anime Style (Beta)

See Preview Images for my Prompts.

Preview Images generate:

LoRA Load over: additional-Networks (work without it)

Model: abyssorangemix2_Hard

VAE: Anything-V3.0.vae.pt

Txt2Img and Upscale settings:

• DPM++ SDE Karras

• Steps: 20

• 480x720

• CFG: 7

• LoRA Strengh: 0.4

• Upscale: 2x

• Denoise: 0.6

• Upscaler: "R-ESRGAN 4x+ Anime6B"

推奨タグ:

SHISOU,high twintails,blue clothes, shaded face, hair ornaments,puffy sleeves, tawny hair,

Yol Briar Ibara-Hime LoRA

This LoRA is designed to summon the beautiful assassin, Yor Briar, the Ibra-Hime.

When combined with the separately offered LoRA of Yor Briar, satisfactory results are often achieved.

You can devise your own weights.

In our sample, we use 0.6 for Ibarahime and 0.2 for Yor.

>>> https://civitai.com/models/33419/yor-briar-yor-forger

If it doesn't work, the target method is to just do the numbers.

And don't stick to the sample prompts.

Thank you for your patience.

ヨル・ブライア いばら姫 LoRA

美しき暗殺者、いばら姫ことヨル・ブライアを召喚するためのLoRAです。

別で提供しているヨル・ブライアのLoRAと組み合わせると満足な結果が得られることが多いです。

重みはそれぞれ工夫してみてください。

サンプルではいばら姫0.6、ヨル0.2としています。

>>> https://civitai.com/models/33419/yor-briar-yor-forger

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

SamplePrompt

Yor Briar, <lora:ibara:0.6> <lora:yor:0.2> 1girl, black hair sideburns:1.8, (slanted eyes:1.4) red eyes, medium breasts, bangs, gold earrings, long locks hair between eyes skindantation :/ angry updo (folded ponytail hair pulled back:1.5), (Folded tied hair:1.5) , (bare_shoulders off shoulder black slit dress black fingerless_gloves, gold hairband with hair flower, gold rose ornament, black thighhighs, high heel boots, collarbone cleavage) fighting stance luxury hotel

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 4292274236,3688654080, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

All the content I make is and will be completely free on CivitAI.
But if you want to support me with a coin, then you can do it here.

Trained on AbyssOrangeMix2 with Clip Skip 2.

Sample images was generated with Luma and AOM2

Recommended embeddings:
1. easynegative
2. bad-hands-5
3. badhandv4

Strength 0.6-0.8 is optimal.

PLEASE CHECK YOUR LORA NAME AFTER DOWNLOAD.

Male Illidari: 1boy, illidari
Female Illidari: 1girl, illidari

Skin color: any, but you can go with purple skin if you want to make Illidan
Eye color: any, but also you may use glowing green eyes, blindfold

Also you may improve it with wings, horns, muscular, glowing eyes, tattoo

Why it's Amy Rose of course!

Helpful positives: pink skin, dress, short dress, hairband, gold bracelet, boots

Jade Harley LoRA

Hello again! This is the third LoRA i've done and am steadily getting better at this process ^^

I've trained it to have her Godtier outfit and dogears, and seems to work best around 0.8-0.9

:)

Hutao (Genshin) | 胡桃（原神）

Base model: anything-v4.5-pruned

Resolution dimensions: 512*768

Trigger word: hutao\(genshin\)

Suggested settings:

• I had CLIP skip 2 on all pics.

• Use VAE if you want to get more vivid images. (I usually use kl-f8-anime2.)

• You can use Hires. fix to improve image quality. (Recommended use R-ESRGAN 4x+ Anime6B Upscaler.)

• Don't use restore faces.

• Set the weight to 0.7

建议：

• 将CLIP skip参数调至2

• 使用kl-f8-anime2 VAE

• 可以使用Hires. fix提升画面分辨率 （Upscaler推荐使用R-ESRGAN 4x+ Anime6B）

• 不要使用restore faces，个人感觉用了后嘴部不像了

• 模型权重调至0.7

Weight: 0.5 - 0.8

Tested dimensions:

• 600*900

Note: This is an early version with several known issues. If you want to view her feet (like me), be patient.

dio:dioxth

kujo:joxsp,stand \(jojo\)

weight:0.8-1

通过分层缓解了画风过拟合的问题

The problem of overfitting of painting style is alleviated through lora block

weight 0.6 is best.

I created this LoRA in response to a request.

selene1, 1girl, selene1, 1girl, selene \(pokemon\), solo, grey eyes, black hair, green shorts, red headwear, beanie, shirt, tied shirt, floral print, short hair, short sleeves, short shorts, striped, yellow shirt, belt, midriff

selene2, 1girl, solo, white shorts, selene \(pokemon\), twin braids, shirt, hat flower, long hair, bangs, red flower, orange shirt, eyelashes, sleeveless, brown hair, brown headwear, grey eyes, sleeveless shirt, short shorts, jewelry





If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

weight 0.6 is best.

This LoRA represents four different outfits for Sakura Miko.

miko1, sakura miko, 1girl, green eyes, solo, ahoge, x hair ornament, pink hair, single thighhigh, sideboob, hairclip, hair bell, floral print, long hair, hair between eyes, one side up, white thighhighs, asymmetrical legwear, wide sleeves

miko2, sakura miko, 1girl, solo, green eyes, long hair, ahoge, black headwear, hair ornament, white shirt, black thighhighs, pink hair, red skirt, plaid skirt, garter straps, collared shirt, hairclip, frills, bangs, hair between eyes, animal hat, frilled skirt, beret, pleated skirt, hair flower, low twintails, neck bell, cat hat, medium breasts, puffy long sleeves, black bow, underbust

miko3, 1girl, sakura miko, solo, green eyes, ahoge, cleavage, pink hair, neck bell, cat hair ornament, short shorts, bangs, hair between eyes, two side up, red shorts, red shirt, medium hair, black choker, pom pom hair ornament, short sleeves, cross-laced clothes, bare shoulders, thigh strap, wrist cuffs, hairclip

miko4, 1girl, hololive idol uniform, sakura miko, idol clothes, green eyes, ahoge, pink hair, solo, x hair ornament, hairclip, single thighhigh, white gloves, long hair, floral print, bell, hair bell, cleavage, pink thighhighs, one side up






If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

weight 0.6 is best.

I created this LoRA in response to a request.

nmm1, nekomusume, nekomusume \(gegege no kitarou 6\), 1girl, purple hair, yellow eyes, solo, red dress, red choker, hair bow, pointy ears, short hair, white shirt, long sleeves, red bow, bangs, collarbone, brooch, jewelry






If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

A Lora of Wicke from pokemon sun and moon.

The tags are: turtleneck sweater, capelet, skirt, glasses.

Unfortunately the white strap over her chest doesn't go away if you remove the capelet.

Use Lora weight of 0.7.

weight 0.6 is best.

I created this LoRA in response to a request.

The tails are very random.

You can either remove the tails from the prompts so they don't appear at all, or you need to adjust the tails with sketch and inpaint.

neferpitou1, cat ears, (cat tail:0.8), solo, short hair, red eyes, long sleeves, hair between eyes, :3, curly hair, striped, cat girl, buttons, striped shorts




If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

weight is 0.6 best.

I created this LoRA in response to a request.

When I put the dataset together, I had a very difficult time creating it because the images in the dataset were mostly dark backgrounds with a lot of colorful images.

1girl, solo, witch hat, long hair, large breasts, detached sleeves, garter straps, sideboob, hair between eyes, cleavage cutout, bare shoulders, black thighhighs, belt, grey hair, dress, pelvic curtain, side slit, vertical stripes,

I apologize if you feel that this LoRA does not resemble the original character.

If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

weight 0.6 ~ 0.7 is best.

venus1, 1girl, solo, sailor senshi uniform, blonde hair, magical girl, blue eyes, orange skirt, elbow gloves, tiara, pleated skirt, hair bow, orange sailor collar, miniskirt, red bow, orange choker, white gloves, very long hair, jewelry, earrings

If you add 1990s (style) to the prompt, you can add an animated effect to the image.






If you want to request a different character or have any questions, please use this Discord.
https://discord.gg/x2nhwavQNF

Kohya GUI, 400 Images, 15 steps each.

yuki is a famous influencer and asian model.

The images have been gathered from instagram.

You are responsible for the images created by you. Unethical usage of this LORA is prohibited.

I strongly advise against uploading any NSFW generated by this model. I warmly welcome you to communicate with me here on how to improve the working method to make the model more realistic. I am currently considering how to generate a better version 3.0. If you have any suggestions, such as choosing different base models, training with more images, increasing the number of training steps, or changing the training size to achieve a more realistic overall effect, please feel free to leave a comment below.

I advice to put this model weight always to 1 for best results.

Look at example images if you are looking for good prompts.

Comment and post your best images, i'm waiting for them!!

建议搭配脸部的lora模型使用，

使用chilloutmix 模型权重大约0.9

使用动漫模型时，权重大约0.9

It is recommended to use with the Lora model of the face

Using the chilloutmix model, the weight is approximately 0.9

When using animation models, the weight is approximately 0.9

请我喝一杯奶茶或者更多？

Can I have a cup of milk tea or more?

This is a New Style lora I have created, it is suitable for cartoon styles. (I have not test this lora with the realistic style yet so I am not sure how well it handles.)

Another reminder I have only tested this lora with camelliamix models, and some other 2d models. Work best with camelliamix models.

I will suggest using 0.6 to 0.8 weight for this Lora.

In the example images I used the same prompt and same Model

Model: camelliamix_v10, AnyPastel

Sampler: I will suggest using Euler a, DPM++SDE Karras and UNIPC.

Prompt: (long braided hair), (bangs), ultra-detailed, 8k, masterpiece, beautiful detailed face, (volumetric lighting), beautiful detailed eyes, (ambient light), realistic shadows, (fullbody:1), 1girl, (cowboy shot), looking at viewer, perfect body, body well:1, leg well:1, arm well:1

邪神ちゃんはわたしがいないとダメだからなあ〜

Please use Hire.Fix

Works under 1.0 weight

Hope you like it.

https://t.me/stdif_kos

The Volkswagen Golf Mk4[3] (or VW Type 1J) is a compact car, the fourth generation of the Volkswagen Golf and the successor to the Volkswagen Golf Mk3. Launched in October 1997 for the 1998 model year, it was the best selling car in Europe in 2001 (though it slipped to second place, behind the Peugeot 206, in 2002).[4]

The Mk4 was a deliberate attempt to take the Volkswagen Golf series further upmarket, with a high-quality interior and higher equipment levels.

It was replaced in 2003 by the Volkswagen Golf Mk5 in European markets. However, manufacturing continued in South America and China for developing markets until 2014.

V2 added more images into to the data set for more diverse facial angles.

Add zoom_layer into the prompt will generate a face in the background:

Add upper_body into the negative prompt if you want the face for avatar only.

Not need to add watermark, username into the negative prompt since it will make no difference in result.

V2 Sample image:

V1 Sample images:(look at here if the images disappear)

白目をむいてガーンってなるLoRAです。

多くのキャラLoRAとの併用ができます。

The LoRA created from the artworks of Kim Jung Gi.

••••••••••••••••••••••••••••••••••••••••••••••••••••••

For optimal results, I highly recommend downloading the additional KimJungGi LoRA from my Hugging Face repositories or download it from here in training data tab. By incorporating it/mixing weights into Additional Networks, you'll be able to achieve even better outcomes.

••••••••••••••••••••••••••••••••••••••••••••••••••••••

To achieve the unique and captivating style of Kim Jung Gi, try using the trigger words "kimjunggi style" or "kimjunggis style". Additionally, consider adding prompts like "sketch," "heavy sketch," "rough sketch," "underground comic," or "concept art". (see example images)

••••••••••••••••••••••••••••••••••••••••••••••••••••••

I've made my very first LoRa. I'm still getting the hang of it, so any advice or recommendations you have would be incredibly valuable to me. And, of course, don't forget to share your artwork here! Thanks for your support and let's have some fun!

••••••••••••••••••••••••••••••••••••••••••••••••••••••

weights example : https://imgur.com/NQXCsJZ

Use to 1 ~ 0.8

Since it was trained with very few images, overfitting will occur.

However, you can try different outfits by writing prompts.

1.Normal

stella_unibell, navel,

2.Playboy bunny

purple_leotard, rabbit_ears,

3.Swimsuit

school_swimsuit, sailor,

Yor Briar (Yor Forger) LoRA

LoRA to summon the beautiful public servant, Yor Briar.

The weight should be around 0.7.

By adding the DynamicPrompt extension, it is possible to randomly create a clerical uniform and a private uniform as shown in the sample prompt.

If it doesn't work, the target method is to just do a few.

And don't get hung up on sample prompts.

Thank you for your patience, it is much appreciated.

ヨル・ブライア (ヨル・フォージャー) LoRA

美しい公務員、ヨル・ブライアを召喚するためのLoRAです。

重みは0.7前後がちょうどよいでしょう。

DynamicPromptのextensionを入れることでサンプルプロンプトのように事務員制服と私服をランダム錬成できます。

うまくいかない場合の対象方法はひたすら数をこなすこと。

そしてサンプルプロンプトにこだわらないことです。

あなたの忍耐に感謝いたします。

New!!

LoRA of Ibara-hime was also released!

いばら姫のLoRAもリリースしました！

https://civitai.com/models/33529?modelVersionId=39830

SamplePrompt

yor briar, <lora:yor:0.7:NP> 1girl, black hair sideburns updo red eyes, medium breasts, bangs, gold earrings smile :> { white open-collared shirt, turquoise vest, turquoise Pencil Skirt, office | long length red sweater_dress, off shoulder, long sleeves collarbone cleavage, dining room }, black pantyhose, white hairband

Negative prompt

badhandv4 bad_prompt EasyNegative

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 2,3, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires steps: 20, Hires upscaler: R-ESRGAN 4x+ Anime6B

Use Checkpoint

使用したチェックポイント

caulkinumV2_AR

https://civitai.com/models/32537/caulkinumv2ar-for-anime-style

Donation

Coffee will increase work efficiency and motivation.

寄付

コーヒーは作業効率を高めモチベーションをアップさせるでしょう。

https://ko-fi.com/489489

Thank you for reading to the end.

I hope one of you will use it.

最後まで読んでくれてありがとうございます。

使用してくれるか方が一人でもいることを祈っています。

我使用anythingv4.5，权重0.8-0.9。

用flower over eye可以召唤出眼睛上的花。使用realistic可以让人物3d化。搭配navel让衣服更好看。blood,  blood on clothes, blood on face可以增加一点鲜艳的色彩。

I use anythingv4.5 with a weight of 0.8-0.9. Trigger words "flower over eye" can summon flowers over the character's eye. Using "realistic" can create a 3D effect on the character. Adding "navel" can make the clothes look better. Trigger words "blood", "blood on clothes", and "blood on face" can add a bit of vivid color.