[2023-04-10]

Recommend options

• LoRA weight 0.6~0.7

• Trigger words

  • base

    • Cure-Spicy

  • Keywords for more detailed costumes:

    • (blue hair,)multicolored hair,((Right side ponytail,side ponytail,))

      blue legwear,dress,gloves,((navy blue pantyhose)), puffy sleeves,white gloves,navy blue footwear,brooch,jewelry,bangs,frills,flower,((blue skirt,long skirt,))high heels,jewelry,

    • Try different combinations of words and coordination!

  • Her hair side and costume are very detailed and difficult to reproduce perfectly!

    We need comrades who can make a higher quality LoRA!

Use animefull based model

• Settings

  • Stable Diffusion WebUI by Automatic1111

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

  • about 20 steps, CFG scale 3.5~6.5

  • kl-f8-anime2.vae

  • CLIP skip = 1

  • Use highres. fix to get higher quality image

    • Upscaler : Latent series

    • Denoising strength 0.50~0.65

LoRA training info

• trained on sd-scripts by kohya_ss. Thank you a lot!

• 109 images of Cure-Spicy

• replaced character feature tags with Cure-Spicy

• resolution : 512x512

• rank=alpha=16

• batch size = 4, 7 epochs trained

This LoRA may be compatible with animefull-based models (ex. AbyssOrangeMix Series...)

This LoRA can generate nsfw images of Cure-Spicy

This LoRA can not generate nsfw images of huwakokokne

p.s. It's not necessary but If you like my LoRAs please consider supporting me on Kofi or order commission there

(LINK ON THE COFFEE MUG ICON ON MY PROFILE)

Please give 5 star reviews by sharing your picture creation if you like it,

• (There's 2 option of version on this release, please experiment both)

• In general This guide basically works for both version, but the experiment ver. need some tweaks)

• Use any checkpoint based on SD1.5

• its 768,768 image

• turn on Hires.fix (or not, please experiment with it yourself)

• Hires steps: 14-17, denoising:04-65 ish.. just try it yourself

• CGE: 5-7

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

• Use 20-ish sampling steps

• Strength: 0.6-1

• I suggest turn off the restore faces check mark for better results of this Lora

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

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 local laws and regulations, and do not infringe other people's reputation, privacy or portrait rights.

I generally do not support the use of this model to generate any adult content. When using this model, users should abide by local laws and regulations and shall not violate any other's right to reputation, privacy or portrait.

I generally do not support using this model to generate adult content. When using this model, users must comply with local laws and regulations and not infringe on others' rights of honor, privacy or publicity.

This model is made by me and has nothing to do with the person who posted it before.

This model is made by me and has nothing to do with the person who posted it before.

This model is made by me, not related to the previous poster.

First LORA but seems to work okay. Best at .6-.8. LORA was trained on ReV Animated but seems to work with a variety of models though some tend to lose her hair feather, works with sfw and nsfw. Some models also seem to need "tan skin" added or it makes her very pale.

female pov for a few scenes

pretty gacha for some

in order of how effective the triggers are:

female pov, pov breasts, pov crotch, cunnilingus, futanari pov, female masturbation, fingered pov, fingering pov

• female pov: female pov alone adds breasts, and makes the pov more feminine

• pov breasts: adds breasts

• pov crotch: adds pov lady parts

• cunnilingus: a face between your lady legs

• futanari pov: feminine body, breasts, and a large manly bits

• female masturbation: pov, gacha, has great results, and terrible ones

• fingered pov: pov of getting fingered, gacha

• fingering pov: pov of doing the fingering, gacha

original description

Since this painter repeatedly drawing the same character hatsune miku many times before, if you input the hatsune miku tag, this will generate quite good results, such as:

However, the lack of art work diversity caused not such good results to other characters (without other model):

I uploaded a wrong file accidentally with this same title on civitai in few days ago(7th April), and removed it in few hours later, if you downloaded it, please remove it and download this one. That file is far less accurate.

If you like my work, please upload the images you made by this model here.

Since this painter repeatedly drawing the same character hatsune miku many times before, if you input the hatsune miku tag, this will generate quite good results, such as:

However, the lack of art work diversity caused not such good results to other characters (without other model):

I uploaded a wrong file accidentally with this same title on civitai in few days ago(7th April), and removed it in few hours later, if you downloaded it, please remove it and download this one. That file is far less accurate.

If you like my work, please upload the images you made by this model here.

Kohya GUI, 150 Images, 15 steaps each, 2 epochs.

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, i also advice to use DMP++ Sampling methods and with a CFG of 7 or similar.

Here's a LoRA specialized at depiction of the Derriere Bagels. Trained on anything-v3, kinda biased towards rear view butt shots but can do other poses as well.

Trained mostly on furries but there's Zack3D compatible version that can do humans as well.

Used e621 tags w/o underscores while training. Putting balls, butt shot, rear view in negatives helps diversifying poses.

Works really well with AnyhtingFurry model, should be compatible with other anything-v3 based models.

original description

SFW LoRa from all shoujo ramune anime characters. LoRa dataset not contain any NSFW images.

trigger words:

yuuko sayama - mother of Chia, works bad, don't use it

chie sayama -

komako semenovich -

tenka adachi -

kiyoshi tachikawa - male protogonist of this title

and "background" for generate title style backgrounds, works bad, but possible to get some results:

strength 0.6-0.8, hiresfix or img2img for best results(it's works for every LoRa).

if you get bad results - try https://civitai.com/models/4669/corneos-7th-heaven-mix this checkpoint and write in prompt position(like "front view, face"), because dataset not only front-face.

Train models isn't easy, crypto donations on some coffee are welcome:

https://plisio.net/donate/pm9qVYci

SFW LoRa from all shoujo ramune anime characters. LoRa dataset not contain any NSFW images.

trigger words:

yuuko sayama - mother of Chia, works bad, don't use it

it doesn't stop -

Komako Semenovich -

tenka adachi -

kiyoshi tachikawa - male protogonist of this title

and "background"for generate title style backgrounds, works bad, but possible to get some results:

strength 0.6-0.8, hiresfix or img2img for best results(it's works for every LoRa).

if you get bad results - try https://civitai.com/models/4669/corneos-7th-heaven-mix this checkpoint and write in prompt position(like "front view, face"), because dataset not only front-face.

Train models isn't easy, crypto donations on some coffee are welcome:

https://plisio.net/donate/pm9qVYci

Why is it called hawthorn cake? Because it is soft and sweet, sour and not greasy~

Eng description below. Please see the photo album below for more examples~

Why does the above example show inpaint: I used Tiled Diffusion, that is, the Multidiffusion plug-in to upscale. The main purpose is to make the picture look better. You can upscale in the normal way, of course, you can also use txt2img to produce pictures.

https://civitai.com/posts/136081 MD前原图在这儿~

Reason why the sample images displayed as inpainting:I'm using the Tiled Diffusion, aka Multidiffusion plugin to upscale, you can of course generate with txt2img——myself did that, and sent the outcome to img2img, then upscaled it lel. https://civitai.com/posts/136081 Original files generated using txt2img here.

这个风格lora里有一个软软的画风，触发token我设置成了child，会让眼睛圆圆的，腮帮肉肉的，总之可爱的（

Other recommended tokens: hanfu, red lips, scenery, east asian architecture, light rays, light particles, sparkle, sliding doors, glowing

Recommended CFG: 7 if there is no Dynamic Thresholding, but I strongly recommend the Dynamic Thresholding plug-in, open 20CFG, everything looks ten thousand times better (?

Recommended resolution: everything is fine, as you can see in the picture, there are all kinds of resolutions, so there is nothing wrong (grass

推荐模型：我是在based64mix上测试的，风格最贴近原版，不过AOM3也不是不行，其他的自己试试也好啦。

I've set the trigger token of the "softer" style in this lora as child, makes the eyes larger, cheeks pinchier, everything cuter!

Recommended tokens: hanfu, red lips, scenery, east asian architecture, light rays, light particles, sparkle, sliding doors, glowing

Recommended CFG: 7 without Dynamic Thresholding Plugin, 14-20 with Dynamic Thresholding.

Recommended resolution: literally anything has been successful on my end, you can see it in the preview pics lel.

Recommended model: I've tested on based64mix, and the outcome is by far the best, however, AOM3、holokukiv2.2、somethingv2_2 could also produce some nice things, feel free to test on other models!

这是一个艺术向hand drawn style模型，虽然命名和示例中又很多校园元素，但是并不代表这是一个校园lora。

Recommended configuration:

• Base model: anything

• Weight: 1.0

• Sampling method: Euler a

• Turn on HD Repair

• Cue Word Correlation (CFG): 8

This is a hand drawn style lora.

“YiWuJiaoYu”Refers to an education policy in China that benefits the people

Recommended configuration:

• Base model: anything

• Weight: 1.0

• Sampling method: Euler a

• Turn on HD Repair

• Cue Word Correlation (CFG): 8

Say hello to Black Jack ni Yoroshiku

is a japaneses medical manga that has a free to use contract, that only requires given regards to the origianl publication. This model use follows the same terms of use.

content using this model on publication give regards to the original source material with examples bellow (Documentation)

"Give My Regards to Black Jack SHUHO SATO", "Give My Regards to Black Jack Shuho Sato"

#Say hello to BlackJack #BlackJackAI

All image scrapped from the free to use material from Tadapic.com . Total of 612image where used. Image where resized to the max size of 512 keeping the original aspect ratio.

https://www.tadapic.com/blackjack.php?p=7#rs. The dataset containt color but mostly black and white mange style images, with all text removed from speach bubbles. These manga style image have the tags "monochrome, greyscale"

Tags where made using stable-diffusion-webui-wd14-tagger then manaual prunned for inaccuracy. Extra tags "blackjack style" was added all image

Training

Model was trained using Kohya LoRA Dreamboothnotebook on a colab pro subsciption

• Basemodel: Anything-v4.5

• Regulation: 800 generic style images generated from Anything-v4.5 with tags from wd14

• Data reperations 2, reg data 1

• Class token blackjack_aesthetic aesthetic

• LoRA Config: Loading network module: networks.lora network args: networks.lora dim set to: 16 networks.lora alpha set to: 8 No LoRA weight loaded.

• LR: 2e-5 (text encoder) 4e-4(Unet) , Consine method

• 20 Epoch Training time 3:28

test

Use 0,4-0,5 weight for more freedom generated image

0,6-0,8 for more detailed feature

Using only Luciaas activating prompt is just fine, but I often using combination of prompt like this : Lucia, Dress, long hair, straight hair, jewelry, earrings, black thighhighs, tassel.

Weight: 0.5 - 0.6

Tested dimensions: W600 * H900

famous idol

Perfectly restored forehead

—————————

There is nothing wrong with running portrait photos, and the feeling is still fully restored, but running half-length or the whole body will collapse;

I added more pictures of various poses when I was practicing, but they came out The effect of running various postures is not as good as the last Jiner...

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 0.6-0.8is optimal.

PLEASE CHECK YOUR LORA NAME AFTER DOWNLOAD.

So, it's trained mostly on Frost Death Knights, but Blood and Unholy works too.

Frost DK: wowdk, glowing blue eyes, skull armor, snow etc.
Unholy DK: wowdk, green skull armor, glowing green eyes, green ghost, undead etc.
Blood DK: wowdk, black skull armor, glowing red eyes, blood rain etc.

Also you may improve it with glowing eyes, skull armor, helmet, fur trim, cape, planted sword, glowing weapon

Please note that it's not the Arthas \ Lich King lora, i will make separate version for him.

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 0.6-0.8is 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

This model's trained with Koikatsu screenshot

Recommneded weight's 0.6~0.7, Sampling steps 20+, Skip Clip 2, cfg 5~11, DPM++ 2M Karras / SDE Karras

Clara main clothes tag: pravda school uniform, green jacket, turtleneck, red shirt, black skirt, pleated skirt, long sleeves, black socks, black footwear, loafers

Works great with anyhentai_18, corneos7thHeavenMix_v2,

anything v4.5, grapefruitHentaiModel_grapefruitv4 etc. models but orange's not recommended. (I'll try fix or research this part)

Trained on 50 photos, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6 on ChillOutMix. lightweight LoRA, works on any generator (pixai, easydiffusion,stablediffusion etc). share your results. enjoy.

This model's trained with Koikatsu screenshot

Recommneded weight's 0.6~0.7, Sampling steps 20+, Skip Clip 2, cfg 5~11, DPM++ 2M Karras / SDE Karras

Nonna main clothes tag: pravda school uniform, green jacket, turtleneck, red shirt, black skirt, pleated skirt, long sleeves, black socks, black footwear, loafers

Nonna detail tag: girls und panzer, nonna, swept bangs

Works great with anyhentai_18, corneos7thHeavenMix_v2,

anything v4.5, grapefruitHentaiModel_grapefruitv4 etc. models but orange's not recommended. (I'll try fix or research this part)

A simple LoRA that implements a super crop top for your female AI character.

Works best weighted around 0.6 or 0.7, but that's entirely up to you :)

Call out prompt tag is SuperCropTop

Any issues please let me know so i can update in future.

I am very pleased if you comment, I gladly accept new ideas, to create new faces.

Based on over 100 images, with backgrounds and movie descents. However, the face is also useable in other contexts.

Visiting my site
https://promptai.altervista.org

No.1xx are some people I don't know at all, I may put them in this series because I am too lazy to change the name.

No.2xx will probably be my favorite celebrity or blogger.

The strength of lora is recommended to be 0.6-0.8. Chilloutmix produces a picture, see the schematic diagram for the entry, and I used the EasyNegative.pt file for the negative.

The material comes from the network wallpaper.

Public channel: t.me/dawn66666666

Jill Stingray from the great Va-11 Hall-a

a personal favourite of mine that visual novel.

Because I forgot to set the trigger word, it caused the characters to blend together, and the drawing was the same as the lottery

Was playing her today for current event an thought she cute make a lora, so i did.

Redone with more screen caps, found her in 3 locations and took a bunch of pics

Use strength 0.2-0.5 for 2D and 0.55-0.8 for 3D .

share ur gens would love to see what comes of her have fun

This is my first experience with the LoRa model.
Work with Deliberate checkpoint, can generate characters with the face of famous Russian politician Vladimir Zhirinovsky.
Trigger words is "Zhirinovskiy", "Zhirik"

A model that adds the style of old memes, such as trollface.

Add ",old_memes style " to your prompt to use it.

JAV model from early 2000, now retired.

Rename the downloaded Lora file to "AkiraWatase.ckpt" and then place it in the Lora folder.

To use, add this anywhere in the prompt:

AkiraWatase

Weights 0.65-1.00 recommended, see examples. This model has been sufficiently trained to produce likeness up to medium body shots.

Leave a review if you like it and checkout my other models.

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

MS Real is a new collection I'll be doing. Instead of Anime images like I do normally in "Murky's" MS Real will be focussed on realistic models. So far as I tested it works on anime models too.

Trained with 90 images with mirrored versions and 768x768.

Use 512x and hires fix for better results

Tags you might need are

1girl, pussy, solo, anus, lying, ass, spread legs, on back, legs up

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

Known Problems

-Too many fingers or too less. You know the deal

Discord server feel free to drop and say hi ^^

Remember to use the trigger word water. Use LORA weight 1 and Hires fix with latent bicubic antialiased, 20 steps, 0.4-0.6 denoising strength, and 1.25x upscale. I've used ChilloutMix, RL02Mix, and KanPiroMix to generate pictures.

Trained with 394 images with 10 max epoch and 1 repeat.

Triggers with bangchan and bang-chan weight 5.1 works best with most models. Usedany realistic modelfor example images.

The model is in the original Ghibli style and can be used with other Loras in the generative animation style which can create fantastic scenes and buildings。

Download counterfeit-v2.5 model before use

If you like my model, please share it with others or follow me. I will update better models for environmental concept artists

v2.0 The light and shadow effect of the structure is optimized to be more friendly to the architectural expression. Please adjust Tag to avoid unreasonable building structure

This model is not for commercial use, only for exchange learning

Optimal values:0-6-0.8

Trained on the NAI model

Enjoy!

If you like the model, you are free to share your creations, I really like seeing everyone's results! :)

A Lora in the style of Rizdraws.

I thought I'd try my hand at a style Lora.

If you are using it on its own use a weight of 1, if you are using it with another lora use a weight of 0.7

Not perfect, will remake in the coming days.

Hello everyone, please allow me to introduce my new model puresexy , through this model you can get a girl with an innocent face and sexy body, the model is used without any restrictions, by increasing the weight to enhance the style, 0.1 ~ 1 are OK, it is recommended to use hiresfix to get more details, this model can be used with any base model, enjoy!

Base model below has been tested:

Anything v4

perfectworld v2

chilloutmix ni

Hello everyone, please allow me to introduce to you the new model puresexy I made, through this model you can get a girl with a pure face and a sexy body, there is no limit to the use of the model, and the style can be strengthened by increasing the weight, 0.1~1 is all ok, recommend using hiresfix for more details, this model can be used with any base model, have fun!

original description

Trained on nai with 78 pictures
Good results with 0.8-1 weigths.

Comissioned by Auxil

Reviews are really appreciated, i love to see the community use my work, that's why I share it.
If you like my work, you can tip me here.

Got a specific request ? I'm open for comission on my fiverr gig *! If you provide enough data, OCs are accepted

*Any paid request will not be presented on this page, unless you give me permission to.

Trained on nai with 78 pictures
Good results with 0.8-1 weigths.

Comissioned by Auxil

Reviews are really appreciated, i love to see the community use my work, that's why I share it.
If you like my work, you can tip me here.

Got a specific request ? I'm open for comission on myfiverr gig *! If you provide enough data, OCs are accepted

*Any paid request will not be presented on this page, unless you give me permission to.

Model Introduction: This LoRa model is inspired by the character Vladilena Milize from the anime 86 - Eighty Six. The model is trained based on the anything v4.5 base model and has been tested to exhibit good versatility. However, using the same type of model for generation may yield even better results. This is the first version of the model, and the creator will continuously test, optimize, and update it. If you encounter any issues during usage, please feel free to provide feedback to the creator.

Recommended Weight Range: 0.7-0.9

Training Datasets and Trigger Words: This model utilizes three separate training sets, targeting the character's facial features, hat features, and clothing features. The corresponding trigger words are as follows:

1. Facial Features: face

2. Hat Features: lenahat

3. Clothing Features: Vladilena (note the case sensitivity)

Usage Restrictions: The use of this model and the images generated by it is strictly prohibited for any form of commercial purposes. Please comply with relevant regulations and use responsibly.

Enjoy using the model!

ANIME LORA Trained on NAI.

Two-dimensional lora based on NAI training.

"hair between eyes" and "white _hair" would be better.

Use "hair between eyes" and "white _hair" for better results.

you can try ((hood up)).

You can try ((hood up)) to put a hood on Xiaobai.

Ji, the instructor of campbuddy, knows everything.

Ringleader Scarecrow of Sangvis Ferri from Girls' Frontline

Keywords:
ScarecrowSangvis
ScarecrowSangvisMaskless
ScarecrowSangvisSchool (Mouth Mask)

Recommend 0.8-1 strength

If you've enjoyed my models, consider leaving a tip.
https://ko-fi.com/leontalon

Intro

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

Version 1.0

It works well with all kinds of anime models, such as Anything Family, SoftPastel,Yaoi, Pastel2D etc.

You can simply use this Lora to draw your boy a blue hair and a deep blue hat.

Trigger word "mizuki \(arknights\)" is essential,"mizuki \(outfit\)" is for the umbrella or more detailed clothes.

Please tell me if you have any suggestions. With much appreciation.

original description

GOBUTA ゴブタ (That Time I Got Reincarnated as a Slime)

Fufufu~ this is the time where you've all been waiting for! The time for our favorite Hobgoblin to finally have his own stage! Introducing......... drum rolls THE GOBUTA LORA

The LORA has 3 versions, 20, 10 and 13 epochs. I personally like the 13 epochs as it gives better outcome and not too overfitted so Gobuta can wear other clothes. You can get the other versions for free in my pixiv fanbox!

Subscribe to my Fanbox to see my latest LORA and artwork.

Recommended Lora strength : 0.4-0.9

SWEET SPOTS : 0.6-0.8

This is a free lora that I made where you can summon Gobuta and give him cute costumes. Here are the trigger words to use them

ふふふ～待ちに待ったこの時！私たちの大好きなホブゴブリンがついに彼自身のステージを持つ時が来ました!紹介...... ドラムロール THE GOBUTA LORA

LORA には、20、10、および 13 エポックの 3 つのバージョンがあります。個人的には 13 エポックが好きです

私のpixivファンボックスで他のバージョンを無料で入手できます！

推奨ローラ強度: 0.4-0.9

スイートスポット : 0.6-0.8

ゴブタを召喚してかわいい衣装を着せる無料のロラです。それらを使用するためのトリガーワードは次のとおりです

1) SCHOOL UNIFORM = Gobuta, a cartoon character with a green face, wearing black GOB_uniform, purple mohawk

2) ORIGINAL OUTFIT = Gobuta, a cartoon character with a green face and purple mohawk, wearing GBToutfit

3) WHITE SHIRT = Gobuta, a cartoon character with a green face and a white shirt, purple mohawk

4) SHIRTLESS = Gobuta, a cartoon character with a green face and purple mohawk, shirtless

5) BATTLE READY = Gobuta, a cartoon character with a green face and purple mohawk, wearing GBToutfit, carrying a mace weapon, brown pants

GOBUTA ゴブタ (That Time I Got Reincarnated as a Slime)

Fufufu~ this is the time where you've all been waiting for! The time for our favorite Hobgoblin to finally have his own stage! Introducing......... drum rolls THE GOBUTA LORA

The LORA has 3 versions, 20, 10 and 13 epochs. I personally like the 13 epochs as it gives better outcome and not too overfitted so Gobuta can wear other clothes. You can get the other versions for free inmy pixiv fanbox!

Subscribe to my Fanbox to see my latest LORA and artwork.

Recommended Lora strength : 0.4-0.9

SWEET SPOTS : 0.6-0.8

This is a free lora that I made where you can summon Gobuta and give him cute costumes. Here are the trigger words to use them

ふふふ～待ちに待ったこの時！私たちの大好きなホブゴブリンがついに彼自身のステージを持つ時が来ました!紹介...... drum roll THE GOBUTA LORA

LORA comes in three versions: 20, 10, and 13 epochs. I personally like 13 epoch

You can get other versions for free in my pixiv fanbox!

Recommended roller strength: 0.4-0.9

Sweet spot : 0.6-0.8

A free Lola that summons Gobuta and dresses her in cute costumes. Trigger words for using them are

1) SCHOOL UNIFORM = Gobuta, a cartoon character with a green face, wearing black GOB_uniform, purple mohawk

2) ORIGINAL OUTFIT = Gobuta, a cartoon character with a green face and purple mohawk, wearing GBToutfit

3) WHITE SHIRT = Gobuta, a cartoon character with a green face and a white shirt, purple mohawk

4) SHIRTLESS = Gobuta, a cartoon character with a green face and purple mohawk, shirtless

5) BATTLE READY = Gobuta, a cartoon character with a green face and purple mohawk, wearing GBToutfit, carrying a mace weapon, brown pants

Miss Sara Bellum

Please, I'll appreciate your support on my Patreonandkofi.

Here is a Form you can request me Lora there (for Free too)

https://forms.gle/h9iuLCCnaEs6W9ZB8

If you like our model, please buy me a coffee and supporting for future project.
https://ko-fi.com/apsrcevo
Commercial customization LoRA project are welcome.

After download this model, try to share with us about your finding.
Again, not for commercial use and she is not a existing person.

Please use ChilloutMix and based on SD1.5

• 768,768 image

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

• Use 18 sampling steps

• Strength: 0.7-0.8

• Hires. fix: on

• Hires step: 18

• Sampling step: 18-20

• Denoising: 0.7

Failure armor works

Look like one of my favorite AV actress and already retire.
After download this model, try to share with us about your finding.
Again, not for commercial use and she is not a existing person.

Please use ChilloutMix and based on SD1.5

• 768,768 image

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

• Use 20 sampling steps

• Strength: 0.8-0.9

I am very pleased if you comment, I gladly accept new ideas, to create new faces.

Made based on 50 images, with SD 1.5. Suitable for all models suffice SD 1.5

Realistic Tokai Teio.

This model is recommended for use with ChilloutMix.

some original character, 43 images used

weights: ~0.8

models used for post: grapefruit

Lora for text2img to create cocktails ideally with the right color and glassware.

Trained on images with tags "a {cocktail name} made with {list of ingredients} in a {glass}"

Weight: 0.4 - 0.8

Tested dimensions:

• W600 * H900

• W800 * H800

Kohei Makino from Idoly Pride

Trained on 32 images with custom tags. To use you must use "hsien-ko" or "leilei" in your prompt. I did this so that you could hypothetically have other characters in your image.

There is the issue of her claws. My intention was to be able to generate her trying to do absurd things with her claw weapons. Unfortunately they can't be consistently generated from what I can see. If you add "hands, hand" to your negative prompt, you can get one or two in a batch. All the images I trained that had hands in them, were tagged with "hands".

I had the best luck turning the weight of the Lora to 0.5 to get a good pose and then setting it back to 1 and inpainting.

Character owned by Capcom - I assume no rights of said character.

the checkpoint used is not posted yet.

Would appreciate images, feedback, requests, reviews, and of course comments in the comments. Please check out my other models if you like this one.

it's slightly overtrained but shouldnt be too bad

Kainé from Nier Gestalt/Replicant. probably not better than Lykon's Lora - Kaine (NieR Replicant) LoRA | Stable Diffusion LORA | Civitai

(forgot he made this one lol).

weights 0.6 - 1

to changer her outfit put her outfit tags in the negative prompt and use lower weights

recommended_tags: negligee, elbow_gloves/gloves, bandages, bandaged_leg, hair_flower, hair_ornament, braid

Lyykon's tags might be better,

optional_tags: yellow_eyes, grey_eyes, purple_eyes, medium_breasts, large_breasts one boot, high_heels

the boot is inconsistent, and it may be better to not specify breast size. putting the Nier tag might make it take from 2b/automata ascetics. can gen futa but might struggle depending on the model, since none were in this dataset. if wanting to do futa, this Lora might help Large Futanari LoRA | Stable Diffusion LORA | Citizens

and/ or this one Futa Bulge | Stable Diffusion LORA | civitai

currently don't have the model used posted, but will probably post later or tomorrow

Some of my LoRA's stopped working on the latest Auto1111 version for some reason. Seems to be an issue with LoCon LoRA's. To fix it all you need to do is install this extension then you can use them as normal again!

The Rito species as depicted in Breath of the Wild!

Use tags "rito" and "winged-arms".

This is intended as a Rito species lora, but it can do a few specific Rito to varying degrees of success. It's not perfect for that though, so don't leave a bad review if you can't get a perfect Kass!

Characters trained in this include:
Kass, Revali, Teba, Saki, Amali, Cecili

If going for one of those 6 characters, adding their body color and gender to the prompt can help. for exable, (Revali male dark_blue_body) will be more consistent than just Revali.
Not including "rito" in the prompt and just the character's name can be more consistent as well.

Works well with most models I have tested it with. Have fun!

original description

¿Looking for Jihyo nsfw version?

**Important read**

This model is a early version, Gradually I will add more pictures and updating the model until make it better and more complete

Hello here some settings that can help to get some good results

Recommended resolution: 512x768

• lora weight: 0.70-0.85

• Simpling method: DPM++ SDE Karras

• Steps: 30

• Model: "Chilloutmix_NiPrunedFp32Fix" or "Chilloutmix"

• CFG: 7

You can use some trigger words like: twjihyo, jihyoo

also you can use: "blonde hair", "brown hair" to get some variations

High res is not very recommended but you can use it

I really recommend ControlNet its a big game changer

¿Looking for Jihyo nsfw version?

**Important read**

This model is a early version, Gradually I will add more pictures and updating the model until make it better and more complete

Hello here some settings that can help to get some good results

Recommended resolution: 512x768

• lora weight: 0.70-0.85

• Simplifying method: DPM++ SDE Karras

• Steps: 30

• Model: "Chilloutmix_NiPrunedFp32Fix" or "Chilloutmix"

• CFG: 7

You can use some trigger words like: twjihyo, jihyoo

also you can use: "blonde hair", "brown hair" to get some variations

High res is not very recommended but you can use it

I really recommend ControlNet its a big game changer

original description

A weight of 0.7 ~ 0.8 works well.

I created this LoRA in response to a request.

titania1, 1girl, purple eyes, solo, very long hair, large breasts, black hair, cleavage, bare shoulders, detached sleeves, strapless dress, black dress, long dress, corset

A weight of 0.7 ~ 0.8 works well.

I created this LoRA in response to a request.

titania1, 1girl, purple eyes, solo, very long hair, large breasts, black hair, cleavage, bare shoulders, detached sleeves, strapless dress, black dress, long dress, corset

cetusMix_cetusVersion3

Trained on 45 photos, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6 on ChillOutMix. lightweight LoRA, works on any generator (pixai, easydiffusion,stablediffusion etc)

This is my first LoRA so be aware it may not functioning very well.

Somehow it works better with lower CFG scale(around 5) and I have no idea why this happens. :(

recommended strength is 0.6-0.9

IronCatLoRA #4 - Fate/stay night: UBW Style

Making LoRA's is a thing I do for fun. So if you like what I do? You can help me out by just using the LoRA and leaving a review!

So I like the Fate/ series a lot, I think it must be one of my most favorites' series/shows in anime (Light novels/VN's included). My first LoRA ever made was about one of my most favorites' character, Artoria Pendragon (Saber). So I was experimenting with making a style LoRA based of Fate/stay night: Unlimited Blade Works, since I also like the style of the show a lot. I think I've come to a point after multiple training sessions and testing with big image grids. That this LoRA does its job fairly well.

The style doesn't apply equally good to every checkpoint, but I think the style applies really well to YesMix. All off the current images I've uploaded are created with YesMix, with an exception of the example grids of course, they contain multiple checkpoints. I've offset the LoRA weight, so you can just use a weight of 1.0. This will in my vision give you the nicest possible style possible that will try to resemble the style of Fate/stay night: Unlimited Blade Works. No trigger word is needed, you just need to apply the LoRA.

In order to make it easy for you to see what the actual style changes, I've uploaded multiple image grids, that display multiple checkpoint and multiple concepts. This should hopefully give you an idea what the style on average will look like. It will of course depend on what checkpoint you use, so be sure to check the image grids out to get an idea. The columns that have no text above them, is the generation result from the checkpoint without this LoRA applied. I think for example the city, cityscape, sky styling carries really nice over from the show, also including night is really nice. All character images I've uploaded use the combination of those tags. I also think that trees and grass from the show carry over kinda nice. Feel free to experiment, I've trained this LoRA on all types of concept from the show, even food...

It's also possible to generate characters that occur in Fate/stay night: Unlimited Blade Works. The main characters are easier to generate though, the side characters are a bit harder to generate, but possible. Take a look at my character images I've uploaded to see how you can prompt them. You need to be specific in who the character is, sometimes increasing the weight of the character name might also help. The point of this LoRA is not to generate the characters, it's to apply the style of Fate/stay night: Unlimited Blade Works. I just noticed that's it's possible to generate the characters fairly well, so I've included examples. But you could combine my Artoria Pendragon LoRA with this LoRA in your prompting, to from what I've seen make really nice images, especially on YesMix. I've made an extra image post, with examples on how my Artoria LoRA looks with this style LoRA using YesMix.

original description

It can create very large flaccid penis, which are more realistic, but animations can also be highly versatile, with a personal weight habit of 0.8

可以制作出非常大的阴茎，偏写实但是动画也可以使用泛用性很高，个人权重习惯使用0.8

It can create very large flaccid penis, which are more realistic, but animations can also be highly versatile, with a personal weight habit of 0.8

It can make a very large penis, which is more realistic, but the animation can also be used with high versatility, and the personal weight is used to 0.8

This is a LoRA for creating Shutter Shades.

Requires higher weight to generate the correct look. Example images used weight 1.

If you enjoyed this LoRA, please consider leaving a review. Your feedback helps develop better content. More Jiwenji Creations

Versioning history:

major.minor.patch

major indicates a change in at least one of training data, tagging, or LoRA generation methodology. You can use the prompt from any image with the same major version and it will operate under similar conditions.
minor indicates nothing important unless indicated in release notes

patch indicates the number of training epochs, default is 1

A lora based on an excellent and amazing artist had

If you like his stuff please,please,pleaseconsider supporting him, its the only reason why we have this amazing style!

◆https://nanananafuta.fanbox.cc/

Trained with 100 images

Censored due to oppai loli

WE LIVE IN A SOCIETY

Image info

• Model used abyssorange2hard

• bad hands embedding

• bad artist embedding

• bad prompt embedding

• deep negative embedding

• Lollypop upscaler

• Paizuri Lora (thanks to @ekune)

Known Issues

• None that I have found

Other notes

• This lora generates massiveproportions really well with almost perfect nipples every time!

• Works very well with other loras

• avoid portrait resolution (512x768) if you want maximum thiccness

Training settings?

The most important part of training is getting a good dataset!

Training settings can be found by installing the additional networks extension in auto1111

Extensions -> available -> load from -> additional networks

Restart webui

place lora into stable-diffusion-webui\extensions\sd-webui-additional-networks\models\lora

To view the training settings

Go to additional network tab -> choose model -> training parameters

Come back soon, there is a succubus in the village (well, the iPhone can only upload zip files after trying for a long time, if someone can't use it, I will change to the computer next time)

Recommend Weight: 0.3-0.6 (the bigger the more unreal)

Effect problem: realistic spells will obviously weaken the wings and tail, and slightly weaken the lewd lines. You can add keywords such as wings and tattoo or add descriptions

p>

Watermark problem: The appearance of the devil is strongly related to the watermark of the fan website, and the probability of occurrence is extremely high, but if the negative watermark is too high, it will affect the quality of the picture and even cut it on the head of the lewd tattoo. It is recommended to allow the watermark to appear and then Retouching or cropping

Interpretation of keywords and world view settings (chattering)

1girl: Hidden form, but horns often transform into animal ears Or bun, and the waist will be 20% thinner than ordinary thin women, and the suspected ribs and internal organs are less or smaller than humans

Village_Succubus: demon-like horns, wings, tails, eyes sometimes change color, The limbs sometimes have the pattern of long gloves and stockings, or the pattern of scales and spells

obscenity_tattoo: the manifestation of the magic power of the succubus, which is characterized by the obscenity pattern in the shape of a heart, the stronger the magic power of the succubus , the more love, the bigger and more refined, some succubi don't like to show lewd lines, and some succubi only show lewd lines and don't spread their wings

(Actually, I can't handle wings and tails, obviously training They are both wings and negative. I also try to exclude single_wing. It is a mess. Occasionally it will automatically turn into head_wings but it can still be seen. The position recognition of the sideways and the back is also outrageous. I guess I need special training or I have time to learn controlnet and try... ...

I didn’t let him stabilize the most difficult tail front, and occasionally confused with lewd lines, so I had to give up temporarily, I still want a love tail.

The lewd pattern is quite successful, the real model is too ugly to run out with all kinds of tattoos, doesn’t it look intrusive with a grimace and animal head, so I also specially practiced the love tattoo)

Never forgetti. Works at weight 1. Made using my Lora guide.

Yoshida Hirofumi (Chainsaw Man)

Triggers with yoshida hirofumi, weight 1 works best with most models.

Used Cetus-mix for example images.

It was really a journey to make this LoRA. For such an easy design, it was actually complicated to make it work. It took me 13 tries to get it close to perfection, but well, I think it was worth it.

Usage

General tags: yoshida_csm works fine already, but you can use 1boy, black hair, black eyes, mole under mouth, short hair, crossed bangs if necessary.

For Yoshida's gakuran, just use gakuran, buttons, black pants, sometimes you need to put green pants in the negative prompt (?).

For Yoshida's casual clothes, use official alternate costume, open jacket, white jacket, black pants.

V2.0 CHANGELOGS

• More stability and quality than v1.0.

• Changed the training parameters.

• More flexibility, can easily change clothes.

TRAINING PARAMETERS

• Trained on the old version of kohya-LoRA-dreambooth on Google Colab.

• Trained on a set of 35 images.

  • 10 epochs, 2 batch.

  • 8 repeats

• 0.0001 LR + 0.0001 unet_LR + 1e-05 text_encoder_LR.

• cosine_with_restars scheduler, 0 lr_warmup_steps.

• 64.0 net_alpha, 128 net_dim.

• Training resolution 896x896.

• Model: NAI Model (animefull-final-pruned.ckpt), no VAE.

• 225 max_token_length.

• fp16 mixed_precision.

April 15th is the birthday of the cutest and most edible Zheng Chocho in Wuyinfang

The training set has been uploaded, please contact me if you need the original PSD, jpg and txt files, station B: Yu Lingming; private message me, temporarily only provide to the official, other lora trainers or painters who think I have used your work (actually so far I don’t even have any secondary works)

The training Data has been uploaded

Prompt word: zhiqiangqiang,1girl,

Additional optional prompt words: hanfu, twintails, hair bun

I think it's a bit strange to wear fishnet stockings and boots with the national style, if you think so, you can use barefoot

You can add white background and simple background to negative prompts

It is recommended to use the DPM ++2M or DPM ++SDE algorithm for running the map

The weight can be adjusted between 0.7-1 according to different models;

It can be used with hire fix, please select R-ESRGAN 4x+Anime6B as the zoom algorithm

The first training process used 9 official 3D models, 3 official bust portraits of different sizes, and 1 piece of my own work as training materials

The second training process used 14 AI-generated and modified images, + 2 official images for correction

The copyright belongs to Wuyinfang/Yuehua Entertainment.

Commercial use is strictly prohibited, public dissemination of nsfw images generated by this lora and any behavior that damages the right of reputation is strictly prohibited. If there is anything inappropriate, please contact me to delete this Lora

Prompt words: zhiqiangqiang, 1girl,

Additional optional prompt words: hanfu, twintails, hair bun , barefoot

Additional optional negative prompt words: white background , simple background

It is recommended to use DPM ++2M or DPM ++SDE algorithm. The weight can be adjusted between 0.7-1 according to different models;

It is recommended to use with hire fix, please choose R-ESRGAN 4x+Anime6B as the zoom algorithm

The first training process used 12 Official images and an own works as training Data.

The second training process used 14 images generated by AI and modified, and 2 Official images for correction.

Copyright belongs to Wuyin or Yuehua Entertainment.

Commercial use is strictly prohibited, and public dissemination of nsfw images generated by this lora and any behavior that damages the reputation of zhiqiangqiang is strictly prohibited. If there is anything wrong, please contact me to delete this Lora

This is a LoRA for creating Barista Girls. Get a buzz and a rush.

Instructions:

Put "Barista girl" in the prompt. Tested weights 0.75-1.

Some of the example images use the Lactation LoRA to add natural coffee creamer.

If you enjoyed this LoRA, please consider leaving a review. Your feedback helps develop better content. More Jiwenji Creations

Versioning history:

major.minor.patch

major indicates a change in at least one of training data, tagging, or LoRA generation methodology. You can use the prompt from any image with the same major version and it will operate under similar conditions.
minor indicates nothing important unless indicated in release notes

patch indicates the number of training epochs, default is 1

Ghost of Tsushima style, 建议搭配建筑类型LORA一起使用 ,

My favorite collocation-

Perfect World

Chinese style architecture

TW is 'supersailormoon'

Recommended weight of 0.6-1

This model is exclusively trained on the Crystal and Manga version of Super Sailor Moon (Fuku with a nice gradient). Training images were carefully selected to keep the same aesthetic. Works with most anime models. This is a very accurate lora that will generate consistant results!

Use a lower weight to get least of a style, since it does keep a very Crystal style at higher weight. To help with the lower weight, use extra tags such as ''multicolored clothes, sailor senshi uniform, multicolored skirt, hair ornament, red bow, brooch, heart brooch, earrings, crescent, heart choker, heart, miniskirt, very long hair, back bow, crescent earrings, tiara, yellow choker, boots, knee boots, red footwear"

Eternal Sailor Moon version LoRa here

Not trained on NSFW images, not tested for NSFW

Exemples made with the following models :

• Anything V4.5/V5

• Anylora

• Store Bought Gyoza

• AOM3

• Grapefruit

Recommended Embeddings:

• Easynegative

• bad hands v5

This was made using HollowStrawberry's Lora collab. I highly recommend their guide, has it makes making lora a breeze!

Just a simple LoRA, it was trained on 50 images of Fran from FF12, 10 epochs/1000 steps, trained on Anything v3.

It was originally trained as a Fran LoRA but ended up being pretty good at making Viera or Rabbitfolk.

Activate with the keyword Viera, I recommend keeping it between 0.4-0.8.

I've tested it on a few models and it does well with Realistic and Non-Realistic alike.

original description

介绍：

站内其实已经有几套类似的lora了。但要么炼得太老，要么炼出眼疾。跑出的图总感觉怪怪的不太合我胃口，反正对颜控来说很难受就是了。于是我自己昨晚熬夜炼了一套（深更半夜强迫症发作）。这次的Lora主题为日系美少女。是的，有些像她，又有些不像她。所以命名为飞鸟集。训练集内包含三种主要风格：和服、针织毛衣、双人。

由于C站无法自动识别prompt，这次我投放了大量炼出来的原图。如果你对我上传的某张样图比较感兴趣，你可以另存为后用图片信息功能识别，里面自带所有的prompt参数。可以按照既有参数自行调整。具体的效果请看样图。

相关问题、或者渲染出来的优秀作品，都欢迎在评论区留言反馈。

如果您喜欢并想更多地接触到我的作品，可以在我的主页找到我往期的lora。当然，您也可以在huggingface找到我：

https://huggingface.co/BigXiang

使用事项：

注意！一旦下载这个lora，即证明你自愿遵守以下条款：

• 本Lora仅供个人学习交流，禁止任何形式的转载、传播；禁止用于任何商业用途。

• Lora禁止用于从事非法活动，使用时请遵守所在地的法律法规。对于Lora使用者的非法行为，本人概不负责并坚决反对。

• 不鼓励将Lora用于生成NSFW内容。

prompts of sample:

<lora:FB_style-000008:0.75>,keai,keai_\(keai3\),Sony a6600 Mirrorless Camera,masterpiece,highres,photorealistic,photography,cowboy shot,1girl,solo,pale skin,round face,small lips,small breasts,

Negative prompt: bad-picture-chill-75v, negativeembed,watermark,bad eyes,huge eyes,bad anatomy,(bad arms),broken arm,

Steps: 27, Sampler: Euler a, CFG scale: 7.5, Seed: 3895585265, Size: 568x776, Model hash: fc2511737a, Model: 真人_chilloutmix_NiPrunedFp32Fix, Denoising strength: 0.34, Clip skip: 3, ENSD: 31337, Hires upscale: 1.5, Hires upscaler: Latent

EN:

The theme of Lora this time is a Japanese Kawaii girl. Yes, once seen the sample, you'll felt a little familiar and a little unfamiliar. That's why the name is Flying bird(飞鸟). Three main styles are included in the training set: kimono, striped sweater, and double girls.

Since this website can't automatically identify the prompt, this time I put a lot of original pictures. If you are more interested in a certain sample picture I uploaded, you can save it as and then use the picture information function to identify it, which comes with all the parameters of the prompt. You can adjust it by yourself according to the existing parameters. Please see the sample image for the specific effect and parameters.

Related questions, or rendering out excellent work, are welcome to leave feedback in the comments section.

Terms of Use:

Attention! By downloading this lora, you certify your voluntary compliance with the following terms:

• My Lora is only for personal learning and communication.Any form of reproduction and dissemination is prohibited; any commercial use is prohibited.

• Use for illegal activities is prohibited, and please observe the laws and regulations of your location when using Lora. I am not responsible for and strongly oppose any illegal actions by Lora users.

• It's discouraged for generating NSFW content.

 

introduce:

站内其实已经有几套类似的lora了。但要么炼得太老，要么炼出眼疾。跑出的图总感觉怪怪的不太合我胃口，反正对颜控来说很难受就是了。于是我自己昨晚熬夜炼了一套（OCD attack in the middle of the night）。这次的Lora主题为日系美少女。是的，有些像她，又有些不像她。所以命名为飞鸟集。训练集内包含三种主要风格：和服、针织毛衣、double。

由于C站无法自动识别prompt，这次我投放了大量炼出来的原图。If you are more interested in a sample picture I uploaded, you can save it as a picture and use the picture information function to identify it, which comes with all the prompt parameters. It can be adjusted according to the existing parameters.具体的效果请看样图。

For related questions, or excellent rendered works, you are welcome to leave a message in the comment area for feedback.

If you like and want to get in touch with my work more, you can find my past lora on my homepage. Of course, you can also find me at huggingface:

https://huggingface.co/BigXiang

Usage matters:

Notice!一旦下载这个lora，即证明你自愿遵守以下条款：

• 本Lora仅供个人学习交流，Any form of reprinting and dissemination is prohibited; any commercial use is prohibited.

• Lora is prohibited from being used to engage in illegal activities, please abide by the local laws and regulations when using it. I am not responsible for any illegal behavior of Lora users and firmly oppose it.

• The use of Lora for generating NSFW content is discouraged.

prompts of sample:

,keai,keai_\(keai3\),Sony a6600 Mirrorless Camera,masterpiece,highres,photorealistic,photography,cowboy shot,1girl,solo,pale skin,round face,small lips,small breasts,

Negative prompt: bad-picture-chill-75v, negativeembed,watermark,bad eyes,huge eyes,bad anatomy,(bad arms),broken arm,

Steps: 27, Sampler: Euler a, CFG scale: 7.5, Seed: 3895585265, Size: 568x776, Model hash: fc2511737a, Model: 真人_chilloutmix_NiPrunedFp32Fix, Denoising strength: 0.34, Clip skip: 3, ENSD: 31337, Hires upscale: 1.5, Hires upscaler: Latent

IN:

The theme of Lora this time is a Japanese Kawaii girl. Yes, once seen the sample, you'll felt a little familiar and a little unfamiliar. That's why the name is Flying bird(飞鸟). Three main styles are included in the training set: kimono, striped sweater, and double girls.

Since this website can't automatically identify the prompt, this time I put a lot of original pictures. If you are more interested in a certain sample picture I uploaded, you can save it as and then use the picture information function to identify it, which comes with all the parameters of the prompt. You can adjust it by yourself according to the existing parameters. Please see the sample image for the specific effect and parameters.

Related questions, or rendering out excellent work, are welcome to leave feedback in the comments section.

Terms of Use:

Attention! By downloading this lora, you certify your voluntary compliance with the following terms:

• My Lora is only for personal learning and communication.Any form of reproduction and dissemination is prohibited; any commercial use is prohibited.

• Use for illegal activities is prohibited, and please observe the laws and regulations of your location when using Lora. I am not responsible for and strongly oppose any illegal actions by Lora users.

• It's discouraged for generating NSFW content.

A LORA model of Zoya from Path to Nowhere.

0.7 works best for me.

Lora de la vtuber sumisa!!
Este es mi primer lora, por ende no será el mejor. Espero que les guste!

It's a Wolf of the cartoon "Nu, Pogodi! (Well, Just You Wait!)" shows in the save rooms of game The Atomic Heart. The cartoon is a Soviet riff on Tom and Jerry, with the central premise being the Wolf chasing the Hare.

Poison from Street Fighter and Final Fight. Trained on Lykon's AnyLoRA.

Tested at 0.7-0.9 strength. Trigger phrase is "Poison (sf)".

Add "peaked cap", "collar", "white crop top", and "denim shorts" for her typical look. "Toned body" also helps.

Lora trained on 223 images of Yaoyao

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

yaoyaodef, yaoyaornd,

Trained on NAI, gen'd on AOM2 hard

Lora trained on 1283 images of Nahida

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

nahidadef, nahidardand,

Trained on NAI, gen'd on AOM2 hard

trained with chilloutmix_NiPrunedFp32

I'm bad at writing prompts, so please share yours!

Lora trained on 357 images of Collei

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

colleidef, colleirnd,

Trained on NAI, gen'd on AOM2 hard

Lora trained on 1200 images of Lumine

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

luminedef, luminernd

Trained on NAI, gen'd on AOM2 hard

Dawn, but exclusively trained with the design used in the anime.

Trigger: 'Dawn_Anime'.

*With low weights, I recommend adding 'pink skirt', 'blue hair', 'long hair', etc., to consistently get the basic look. To increase the anime vibe, add 'anime screenshot'.

Weight: Between 0.7 / 0.9. (Clothes seem to survive low values, but the eye shape quickly defaults to 'universal anime style', which sucks. As always, use lower numbers for more creative scenarios and include any missing clothing as a positive tag.)

I trained the model with plenty of pictures with no hat, but the thing ended up being bolted to her head. Try adding '((hatless, head uncovered))' as positive tags and '((hat, beanie))' as negative tags, and it may work… sometimes. If you're lucky.

I tested it with different settings like 'working at a restaurant' and 'coolest surfer in Florida' and seems to work, so have fun!

A LoRA for creating images with women having their panties pulled to the side.

This LoRa works very well if you use it for frontal shots, with the rear view it has a little more trouble. Adjust the weight of the LoRA and the trigger phrase accordingly, e.g. (panties to side:1.2)

If you feel like supporting my work, please consider visiting my Patreon to get access to more detailed instructions on how to get the most out of my models: Link.

Recommended LoRA Strength: 0.6-0.8 from front, 0.8-1.0 from behind

Based on UberRealisticPornMerge

If you like this you might also like my other LoRAs.

Lora trained on 249 images of Faruzan

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

faruzandef, faruzanrnd

Trained on NAI, gen'd on AOM2 hard

Lora trained on 1170 images of Jean

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

jeanrnd, jeanfavonian, jeanseabreeze, jeangunnhildr,

Trained on NAI, gen'd on AOM2 hard

Lora trained on 604 images of Sayu

I release all my models for free, if you feel like supporting me and my work you can support me here and buy me a drink.

Prompts:

sayudef, sayurnd

Trained on NAI, gen'd on AOM2 hard

Kumin Tusyori from the anime ‘Chūnibyō Demo Koi ga Shitai!’ is a character that I really appreciate. A tender and shy girl with an incredible personality. The Lora works well in different models and styles, try playing with the weight for better results. Use the keyword for greater precision ‘KuminWaifu’, the character also includes her characteristic uniform, use thhe trigger word if u want that. Remember to upload an image below to the page, that helps me a lot to continue creating Loras!!”

“IMPORTANT!! If you want a specific model of any style or character that you can’t find, contact me on Pixiv ‘@jkanimations’ and let’s discuss the matter.”