Loha for Mysterious Heroine X and her older version XX. It aims to recreate her with several different outfits. Unfortunately, it did not come out as good as I would have liked.

Use weight < 1.0 .

Outfit tags are:

First ascension: mhx,1girl,ahoge,og_ver

Second ascension: mhx,1girl,ahoge,sec_asc

Final ascension: mhx,1girl,ahoge,fin_asc

Swimsuit: double_x,1girl,ahoge,lulu_swimsuit

Heroic Spirit festival: double_x,1girl,ahoge,her_ver

Open mech: double_x,1girl,ahoge,plt_ver

Full mech: fll_mch

Tend to have simple background if inside or outside not prompted

Tend to portrait if not prompted otherwise, was trained with NSFW images

Trained on NAi 100 images × 3 repeats × 10 epochs ÷ 2 batch size = 1500 steps, face bandaids don't work all time, custom character by the artist jourd4n.

This is a lycors\locon you have to use the additional networks extension or the lycoris extension for this to work.

I made a Lora of this character some time ago and when i was browsing this site i saw that her sister has like 4 different once and i couldn't find one for Mio maybe it's just poorly tagged.

I retrained the model with new settings

Keyword is kofune_mio i use the underscore so it recognizes that as one word sometimes if you have similar words it can overlap.

Works on 1.0 and is very flexible.

If you like it leave a comment, like and rate it that would be appreciated.

Have fun generating

Weight:0.9?~1.0

Trigger : aliciamother, horns, red eyes

when use Official Costumes, add Trigger : cross earrings, red dress,clothing cutout, cleavage, underboob cutout, short white sleeves,white gloves

notice

[cross earrings] is poor. You can remove it.

If used another outfit, there is a possibility of color transfer to red.

This character appeared seven years ago. No kidding...

If you'd like to support me and do more:
https://ko-fi.com/edgonkofi - commissions open!
If you're looking for a >>LoRA Making Tutorial<<

Lycoris Couture

This model explores what we can do with Lycoris and fashion design.

weight：1.1~

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

the regular creativeml-openrail-m license isn't applicable for this work

Feel free to tag me if you use this model, I'd enjoy seeing what you do with it twitter(@ihtke)

UPDATED 5/26/23

see about this version

FIRST 4/25/23

Lemme take u in deep short backstory of my inspiration

Blue Moon🥰 from Cyberpunk 2077 the breathtaking character as game as well.

When had i first heard and learnt about US CRACKS..... oy boy(long sigh). I instantly wanted to meet them in the game (ngl i love every single quest that i completed). And eventually i have been taken a chance to meet them that was abrupt+awesome (as now remembering *incoming call Kerry*

Kerry: drop whatever u doin' meet me at the marina, don't make me wait )more spoilers ahead so u better play it by yourself)

Well...(these warm memories still chasing me) lemme tell u of model a little bit.

Fist thing first all the data that i could scrap and i did were game engine screenshots(intensively waving to the nsfw enjoyers) made by me+from net.

(do not frustrating now there's always is a solution see about that version and +nsfw capabilities also depends of models)

Second as per of data lack this training was a challenge one (as any other train in kohya-ss). I've spent literally a whole(~26h.) day only on trainings (in the name of science i suppose...yeah)

To start🤔:

(masterpiece, best quality, detailed:1.4), Blue Moon cyberpunk, 1girl, eyes, lips, spiked collar, blue hair, long hair, black jacket, yellow collar, solo focus, pink baseball cap, leather jacket, denim, looking at viewer, realistic, cyberpunk, head tilt, outdoors, grey eyes, upper body

(masterpiece, best quality, detailed:1.4), Blue Moon cyberpunk, 1girl, \\m/, bangs, simple background, blue hair, blunt bangs, bracelet, breasts, yellow choker, crop top, double bun, double v, double w, jewelry, lips, looking at viewer, see-through skirt, sport bra, small breasts, solo, standing, tank top, v, v over eye, w, wristband, grey eyes

Important note:

issue with

Kiroshi optics

(needs further investigation)

TL;DR SD sometimes can be a real bitch.

When u trying to make pupils of characters more natural and live then SD will always blur them out and contrary when u gonna make them double on each eye he's mostly just giving up. I recommend whether INPAINT(sketch) or Photoshop to make it properly).

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,

可根据实际需求自行调整，一般来说并不需要写全 tags。

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

SINoALICE style locon

Train based on AnyLora model. Training resources include official arts from Shadowverse,Shingeki no Bahamut and Granblue Fantasy.

Strongly recommend weights from 0.6 to 0.8,and enable hires.fix.

No specific trigger words but you may need add "(hand)puppet,doggy gloves"to prompts or negative prompts to decide whether hide them or not:)

Unlimited Blade Works | Weapon LoHA

This is a LoHA Lycoris and you need this Extension for this to work. I'm happy with the results.

Weight can go from 0.5 up to 1.0. I recommend 0.9. ControlNet will streamline full body shots much easier, or more complex poses. Works with almost every Style, every Character and Photorealistic Models.

This is only a BETA. I will add more Swords, Weapons and Guns to this. SD doesn't like to generate Weapons (just like Ranni with 4 Arms). Trying to improve it on every Update. I recommend using these Settings: Sampling Method DPM++ SDE Karras with 30/40 Steps. If you know how to improve this, i'm happy to hear you out.

WARNING: This LoHA does have some problems and hits most of the time 1 out of 10 times. Hands are a big problem. Hand does not connect with the sword and sword does not appear in the right place sometimes. Unsheathing Sword is currently very unstable. Use pointy ears in Neg, if you are using Link Master Sword. Yor Knife gives normals knifes for some reason on my End.

Trigger Words:

Link Master Sword, Lightsaber, Dual Wielding Swords, One Handed Sword, Unsheathing Sword, Scythe, Sword and Shield, Staff, Trident, Planted Sword on Hand, Yor Knife

I want to add these Concepts in the future as well:

Weapon Classes:

Axe, firearms, arc, Knife, hammer, rods, throwing stars, sickle, whips, guns, grenades, shields

Subcategories:

one-handed, Two handed, pistols, guns, machine guns, shotguns, grenade launcher, rocket launcher, sniper rifles, brass knuckles, throwing axes, spears, javelins, arrows, crossbows, Sling, sniper, bazooka, Foal

Variations:

warhammer, tomahawk, short sword, broadsword, rapiers, longsword, dagger, Kukri, katana, tanto, wakizashi, naginata, quarterstaff, war spear, double axe, hatchet, baton, Shield with integrated weapon, Dual wielding of other weapons

How to use LoHA's in auto1111:

• Download LoHA

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

• Select your LoHA and use the triggerword

• Make sure to change the weight if needed (by default it's :1)

• Have Fun!

This is the second time I trained celebrity, please give me star if you like it. You have to try multiple time to get best result. I recommend to use this lycoris with xxmix9realistic_v30 checkpoint model and set lora weight to 1.

Hope you enjoy.

I've made a new version of the LOHA stockings model. In my opinion, compared to LORA, LOHA performs better, has a smaller size, and has less influence on the background, which makes it easier to train. From now on, all my models will be released in the LOHA format. I recommend a weight between 0.4 and 0.7.

我重新制作了一个LOHA版本的丝袜模型。我认为，与LORA相比，LOHA的表现更好，体积更小，并且对背景的影响也更小，使得训练更加轻松。今后，我所有的模型都将以LOHA的形式发布。我推荐在0.4到0.7之间的权重。

私はLOHAバージョンのストッキングモデルを作り直しました。私の考えでは、LORAと比較して、LOHAの性能はより良く、サイズはより小さく、背景への影響も少ないため、トレーニングがより簡単になります。これからは、私のすべてのモデルはLOHAの形式でリリースされます。重みは0.4から0.7の間が推奨されます。

나는 새로운 LOHA 버전의 스타킹 모델을 만들었습니다. 내 생각에는, LORA에 비해 LOHA는 성능이 더 좋고 크기도 더 작으며, 배경에 대한 영향도 더 적어서 훈련하기 더 쉽습니다. 앞으로 모든 모델은 LOHA 형식으로 출시될 것입니다. 가중치는 0.4에서 0.7 사이를 추천합니다.

Version 3 LoCon is out. It is named Accurate and v2 is renamed Stylized.

Accurate tries to replicate Fire Emblem Awakening Portrait Art as accurate as possible.

Stylized tries to replicate Fire Emblem Awakening Portrait Art style in general.

Keep the weight around 0.6 to 0.8.

Version 2 LoCon is now out. It is trained at 512x512 and produces cleaner results like clearer eyes for example.

Keep the weights around 0.6-0.8, I get best results around 0.7.

Upscale it a little to remove any messiness.

Made using Derrian's Easy Training Scripts

I just realized v1 was actually trained at 800x800, I'll retrain it and upload the v2 fix tomorrow.*
This is a LoCon for generating Fire Emblem Awakening, FE13, character portrait art style.

It's still a bit rough so PLEASE review or leave a comment so I can adjust it as needed.

Also, mix this LoCon with the other Fire Emblem Portrait Art Style LoCons I have for some really interesting and fun results!

This was trained at 800x800* but gens 768x768 perfectly fine.

Recommended weights are 0.65-0.8

Remember, this is a STYLE LoCON so it will reproduce a named character's style, not them exactly as they are. Combine this with character LoRAs for best character specific results.

Note: This model is just for fun and still under test; the overall effect is acceptable but may be unstable. Different models have different effects, and it may be better to choose a model close to the game card illustration style.

This is a LoHa model of MLTD (アイドルマスターミリオンライブ! シアターディズ) game card illustration style, and I try to apply multi-resolution noise to enhance the contrast. The model is based on animefull-latest, and the training set consists of 60 MLTD game card illustrations selected by me. The trigger word is MLTD bg style.

The LoHa weight of sample illustrations is 1.0; you can try other weights.

MLTD game card illustration is colorful, in high contrast and saturation, and usually contains effects of light particles and sparkles. You can add tags such as light particles, sparkle, lens flare, bokeh, blurry background into your prompt, and consider using VAE models such as ClearVAE series and kl-f8-anime series to improve the color of the illustration.

——————

请注意：本模型仅供娱乐且仍在测试中，整体效果尚可但可能不稳定。不同的模型有不同的效果，选择接近游戏卡面的模型效果可能会更好。

此模型为 MLTD卡面风格的LoHa模型，并尝试加入了多分辨率噪声以增强画面明暗对比。模型基于animefull-latest训练，训练集由个人选出的60张MLTD卡面组成。触发词为MLTD bg style。

示例图的LoHa权重为1.0，您也可以尝试其他权重值。

MLTD卡面颜色鲜艳、对比度和饱和度高，通常带有光粒和闪光效果。您可以考虑加入light particles, sparkle, lens flare, bokeh, blurry background等tag，并使用ClearVAE系列、kl-f8-anime系列等VAE模型改善出图的色彩。

Metatron, Archangel of the Virtue of Charity from the Seven Heavenly Virtues (Nanatsu no Bitoku).

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

Trigger words:

• metatron - For character.

• metatronoutfit - For main outfit. Helpful tags: "thighhighs, wings"

Experimental (undertrained) trigger words:

• metatronheart - For heart warmer (cheerleader) outfit from mobile game. Add "cheerleader" for best results.

• metatronswimsuit - For swimsuit from anime. Add "bikini" for best results.

The Queen of Zero no Tsukaima.

Enjoy!

Kasumi Yoshizawa from Persona 5 Royal, trained as a LyCORIS network. Retrained my old LoRA with updates to the dataset and using new training techniques introduced over the past few months.

Version 2 is now out!

This is an all-in-one model with a larger dataset that now includes the appearances of Kasumi, Sumire, as well as the metaverse outfit separately tagged for prompting.

Requires extension for use:

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

Recommended tags are:

Kasumi: dskasumi, red hair, long hair, ponytail, red eyes, hair bow

*Can add sidelocks to the negative prompt to better suppress them from appearing

Sumire: dssumire, red hair, long hair, straight hair, red eyes, glasses

Violet: dsviolet, red hair, long hair, ponytail, red eyes, black ribbon, hair ribbon, black leotard, jacket, red gloves, thighhighs, choker, frilled sleeves, silver rose

Check out my HuggingFace repository:

https://huggingface.co/Doctor-Shotgun/ds-LoRA

Check out my Pixiv:

https://www.pixiv.net/en/users/90447503

Classical art, Rembrandt Style.

Use weight 1 and use the new model, Aphrodite v2. It works BEAUTIFULLY with this LoCON

Here is a recommended prompt:

Rembrandt, (distant view:1.3),(award-winning painting:1.3), (8k, best quality:1.3), (realistic painting:1.1), A gorgeous and intricate painting, fine art

After fine art, add tags you want like ocean, forest, portrait, etc.

Have fun!

Please check out my Ko-Fi page if you are interested in supporting me.

LoRA commissions open! -> Fiverr

This is a LoHa LoRA! You need the extension! Why? They do a lot better with multi-concepts.

Trigger words for V1.19:

<lora:sheathedSwordv119:0.8>(sheathed, sword)
I suggest also using "arms up" or "crossed arms" cus if you don't, the character most likely will be generated holding the sheathed sword.

Follow the steps in the Model Version notes to reproduce my images.

Sample images generated from Nothing model.

All negative embeds should be easy to find. and Prompts should be included in the sample images! Click the little icon in the bottom right to see it.

It was trained on 43 512x512 images scraped from the internet and 20 epochs using animefull_final.

Danbooru tags generated by Waifu Diffusion 1.4 Tagger and fine tuned for specific tags.

Trained on Nai using 30 images, NSFW capable

The suggested weight is 0.5 - 0.7

Based on nai

This is a lycors\locon you have to use the additional networks extension or the lycoris extension for this to work.

This is a lycoris model from the female main character from the anime Dreameater Merry which i didn't even watch until the end it was pretty mediocore but she is cute I guess.

Keyword is Merry Nightmare works well on 1.0 and looks pretty good I was impressed by the quality of the pictures and how flexible it was with backgrounds

I used an old dataset with txt documents that i didn't doublecheck and I think this dataset was pruned like a lot. I don't think you will be able to put her into different outfits i tried a bit and everything looked weird. If you want to get rid of her clothes altogether that might be possible didn't try it though be my guest i guess.

If you want it to be retrained write it into the comments and if enough people have the same sentiment I will retrain it.

If you like it leave a comment, like and rate it that would be appreciated

Have fun generating

***BETA RELEASE: Outfits aside from the dress have proven unexpectedly challenging to get right as the mechanical arm tends to merge with the real one. I'll try to tackle this later, but it might take a while. I decided to release as is in the meantime since it can still achieve some decent generations and I haven't seen any other alternative loras yet.

Belial, Sin of Vanity from the Seven Mortal Sins (Sin: Nanatsu no Taizai).

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

Given the nature of the outfits, the model will tend to nudity. (Just a warning, since I'm certain most of you innocently stumbled upon this for scientific reasons only)

Additionally, watch out for the textual inversion negative prompts you use. While they might help, a lot of these might try to "fix" some details that you would likely like to keep. This is specially true for any model with weird details such as this one.

Trigger words:

• belial - For character. Add "glasses, heterochromia, lipstick" for best results.

• belialdress - For dress outfit. Add "ponytail" for best results

• belialoutfit - For main outfit. Add "ponytail" for best results.

Experimental (undertrained) trigger words:

• belialglory - For alternate outfit from game. Add "gold bikini, ponytail" for best results.

• belialsilver - For alternate outfit from game. Add "long hair" for best results.

This is the first time I created lycoris model. So the result may not be get as expected. In that case, please accept my apologies.

Model: basilKorea_v10

Res: 512 x 768

DO NOT check "Restore faces"

also can generate nsfw, for example:
best quality, MenheraBitch, 1girl, solo, <lora:MenheraBitch:1>, ahegao, spread_ass, open clothes, breasts, nipple_piercing, clothed_sex, doggystyle, vaginal_object_insertion, sex, pussy,gaping, female_ejaculation, excessive_pussy_juice, spread_ass

Silica from Sword Art Online, trained as a LyCORIS network.

One pitfall with this character is that her SAO and ALO outfits share the same design, just with a different color - so they share the same tags during training. This means that SD can sometimes get the color wrong and swap the red and blue outfits. You may have to mess with the prompt/neg prompt a bit to avoid RNG.

Requires extension for use:

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

Recommended tags are:

SAO Outfit: silicabase, brown hair, short twintails, red eyes, armor, pleated skirt, thighhighs, bare shoulders, breastplate, hair ribbon

*Try negative prompting "blue clothing" if you're getting the wrong color

ALO Outfit: silicaalo, brown hair, short twintails, red eyes, cat ears, cat tail, armor, pleated skirt, thighhighs, bare shoulders, breastplate, hair ribbon

*Try negative prompting "red clothing" if you're getting the wrong color

Misc Outfits: dssilica, brown hair, short twintails, red eyes, hair ribbon

I chose not to include "fingerless gloves" as a recommended tag because SD plainly sucks at them...

Check out my HuggingFace repository:

https://huggingface.co/Doctor-Shotgun/ds-LoRA

Check out my Pixiv:

https://www.pixiv.net/en/users/90447503