A LORA model trained on Isaac Rossiter.

To use the model place the words "issacrossiter person" in the prompt.

Trained on 6050 steps.

Ayakashi Triangle //ultimate(bad).

trigger words changed(see in model triggerwords if use non v2.0 ultimate)

Trigger words:

works +-fine:

suzukanade

matsurikazamakif

background

works usable:

shirogane

shiroganetrue

lucytsukioka

seigenkazamaki

ponosukeninokuru

small dataset, but usable:

suzumother

masuraosujimori

i don't know how to use it, it's completelly dead:

yayoitoba - drawn suzukanade. WHY?????? just use "yayoi"

souganinokuru - drawn ponosuke just use "souga" it's works

even if i found previously, this one is hard:

matsurikazamakim -

https://civitai.com/models/42630?modelVersionId=49995 Use this One, works perfect

drawn female variant anyway, too small trigger word difference.

i found bad way to get close : <lora:AyakashiTriangleV2:0.7> (kim), 1boy, red hair

but it's bad anyway

https://civitai.com/models/42630?modelVersionId=49995 Use this One, works perfect

What i learn from made that LoRa... DONT MADE BIG LoRa's with many characters.

Solo character if main(>50 images in dataset), or 3-4 if it's just third plane personages.

Why i upload it... because it's generating like 2 days(and manually dataset preparation +1.5days) and i don't want to do it again. I'ts useful for shirogane, lucy, ponosuke and maybe backgrounds(+ old LoRas characters). maybe i made 1 non useful personages as individual LoRas versions later.

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 contained not only front-face.

i focus on true anime style models for fanarts or fanmanga creation, if you want artstation ultra quality arts style this isn't your choise.

Train models isn't easy, crypto donations are welcome and you can reguest character in it:

https://plisio.net/donate/pm9qVYci

Works best at 0.8

Louis de Funès - French actor (1914-1983)

Please provide constructive criticism.

You can support me here to be part of the society (https://ko-fi.com/djo226). Supporters may request model.

Prompt : Louis de Funes LDF , high quality photography, natural lighting, canon eos 5d mark ii, photorealistic, 8k, photo by Annie Leibovitz, detailed skin, ultra realistic, bokeh, sharp features, centered in frame, facing camera, symmetrical face, ideal human, 90mm lens, f16, photography, ultra details, photo <lora:LDF Cruz:0.8>

Negative prompt: (deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime, mutated hands and fingers:1.4), (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, disconnected limbs, mutation, mutated, ugly, disgusting, amputation

Steps: 80, Sampler: DPM++ 2M Karras, CFG scale: 12, Seed: 739720113, Face restoration: GFPGAN, Size: 768x512, Model hash: 81086e2b3f, Model: realisticVisionV20_v20

Face restoration is optional, may provide better result if not activated

The model changes the expressions and the best results were obtained with:
-Dpm adactive
-Dpm ++ 2m karras
Anything that corrects the face and eyes will leave this parrot much better, define the face well: it helps a lot.
The second version is the same as the first with more training = ergo more face distortion, the best results without parrots that fix the face was the 15 epc

earfuck,<lora:ear-20:1> --->  Activacion
saliva (18)
teeth (18)
tongue out (18)
simple background (15)
tail (13)
black hair (12)
erection (12)
ejaculation (12)
clothed female nude male (12)
hand on another's head (12)
rolling eyes (11)
navel (11)
furry (11)
brown hair (11)
collarbone (11)
short hair (11)
cleavage (10)
oral (10)
bangs (10)
completely nude (10)
kneeling (10)
furry female (10)
multiple girls (10)
2girls (10)
tears (9)
english text (9)
colored skin (9)
multicolored hair (9)
pubic hair (9)
standing (9)
grabbing another's hair (8)
hair between eyes (8)
dark skin (8)
blue eyes (8)
male pubic hair (8)
cum in mouth (8)
purple eyes (7)
ahegao (7)
blonde hair (7)
heart (7)
green eyes (7)
sitting (7)
smile (7)
furry male (7)
bare shoulders (7)
grey background (7)
futanari (7)
after fellatio (7)
small breasts (7)
gloves (6)
choker (6)
cat ears (6)
collar (6)
makeup (6)
bdsm (6)
fellatio (6)
earrings (6)
artist name (6)

Ayaka Hizuki from Memories Off 1st

V1.0

Recommend weight: 0.6-0.9

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

2023.4.18 update

V2.0

The art style now more closely resembles the original game.

The probability of generating wings with errors has decreased now.

Now you can get her wearing the school uniform without having to input any prompts.

If you want her to wear everyday clothes, please input "townwear" and"dress" as the prompt word.

If you want her to wear other clothes, please strengthen the weight of the prompt words, such as (((Nurse wearing))).

Recommend weight: 0.9

Recommend Checkpoint: final-prune, anything, mixpro or other anime checkpoint

Recommend negative prompt: own hands together

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

2023.4.19 update

V2.1

Removed the "townwear" prompt word, now only one prompt word "dress" is needed.

The probability of distortion under high weight is reduced now.

The cancellation of regularization reduces the probability of generating low-quality images.

Improved the overall quality and versatility of the model.

I recommend you use a plugin called "CFG scale fix". All the pictures included in this update were taken using this plugin.

Recommend weight: 0.7-0.9

Recommend Checkpoint: anything

For other usage suggestions of this model, please refer to v2.0

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

This LoRA is based on a model created by Stable Diffusion, not a real person.

Trigger word is 1girl. so if you want multiple girls, recommend use controlnet.

My example images used ChilloutMix.(Base model too.)

Example images contains EXIF.

Hope you enjoy, and feedback is always welcome.

If this LoRa has any problems, i will erase this LoRA and no more updates.

If you have any questions or requests, please contact my Twitter account.

You can try some non-human tags, it would be fun.

Such as:

ice, fire, black hole, earth, moon, lightning, tentacles, abstract, eye, yellow theme……

I even input "dragon", and the result is quite interesting.

There's a small problem in my dealing with datesets, which cause the lower part of the picture to be weird sometimes.

你可以试试一些非人类tag，挺有意思的。

比如：

ice, fire, black hole, earth, moon, lightning, tentacles, abstract, eye, yellow theme……

我处理数据集时留了个小问题，所以图片的底层可能有点怪。

Lora trained on 555 images by Ricegnat.

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.

Trained on NAI, gen'd on AOM2 hard.

徽羽摩柯，moke

如果出了不错的图，可以放到评论区，我会很高兴，用来制作下一版。

Genji from overwatch.

Trained on standard SD v1.5. Good results with revAnimated

PuyoPuyo Style Deformation LoRA

Note: If you like it, please give it a STAR MARKS !

,

LoRA for old-fashioned deformed characters.

The image quality may be rough since we are mainly learning characters from old games.

Therefore, it is recommended to include pixel art tags.

,

If it doesn't work, the only way to deal with it is to just do the numbers.

And don't stick to the sample prompts.

Thank you for your patience, it is much appreciated.

,

,

ぷよぷよ風デフォルメ LoRA

※ 気に入っていただけたならぜひ★をつけてください！

,

古風なデフォルメキャラクターのためのLoRAです。

古いゲームのキャラクターをメインに学習しているので画質が荒いことがあります。

そのためピクセルアートタグを含めるのもお勧めです。

,

うまくいかない場合の対処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 (サンプルプロンプト)

<lora:pstyle:0.5> style parody (thick outline:1.5) 1girl Pixel Art Fullbody dynamic pose

Negative prompt
badhandv4 bad_prompt EasyNegative NSFW

Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 8, Seed: 1943526109, Size: 384x512, Model hash: d5f51c6f0a, Model: caulkinumV2_ARc, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+ Anime6B

,

Checkpoints used (使用したチェックポイント)

https://civitai.com/models/32537?modelVersionId=48898

,

,

,

Use Checkpoint (使用したチェックポイント)

caulkinumV2 シリーズ

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

You can use it to highquality illustrations.

クオリティの高いイラストを錬成できるので使ってみてください。

,

,

Donation (寄付)

Coffee will increase work efficiency and motivation.

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。

Cover eyes

CFG Scale 10+ Increase success rates .

Not supported in real-world scenarios .

【Attention】 that if a single keyword does not work, enter it in the form of this sentence

EX: Strips of cloth, covering the eyes,Cover eyes with ribbon,1girl,

<lora:cover _eyes:0.6>,

挡住眼睛

提示词相关性提高到10左右，

暂时不支持现实场景，当然你也可以试试你的模型。

【注意】如果单一关键词不起作用，可以按我这个例子来输入：

EX: Strips of cloth, covering the eyes,Cover eyes with ribbon,1girl,

<lora:cover _eyes:0.6>,

【如果您请我一杯奶茶的话就更好了！】

【Thank you for your support！】

naizuri: rubbing against nothing

This is paizuri, but with breasts too small to cover the penis.

Also trained a bit on paizuri in order to help it understand the difference.

Using prompt editing [naizuri, flat chest:paizuri, large breasts] has shown good results if you want the naizuri effect with very small breasts.

I have created a LoRA for NCT's Jeno.

Usage: results may vary according to the prompts that you will put. I placed the LoRA weight to around 0.9 to 1.0. Play around with the weights to get the best results.

If you like the model please leave a review or share your work!

--

DISCLAIMER: This model is intended for creative use only. Any images generated using this model is the sole responsibility of the user, and I am free from any obligations or consequences that may arise from the use of the generated content.

weight:0.7

model：meinahentai

Eyes will be re trained. Please wait

Why has no one created an actually good food model yet lol?

Please use this with realistic vision, it only really works with that.



How did i get my thumbnail photos so high quality?

- use prompt, mine is "RAW photo, Spring Rolls, <lora:foodphoto:1> foodphoto, dslr, soft lighting, high quality, film grain, Fujifilm XT"

- find the image you like, enter the seed you like back into the seed box

- switch on highres.fix with a denoising strength of 0.55

- then send it to exrtas on your upscaler of choice

• boom

Face test

Hi this is my first lora, so shes not perfect but I am a potato girl and learning. Miss blue is my main character I play on a modded skyrim game, shes got a special place with me so I try to make her whenever I can create and wanted to make some sexy AI art of her, so sharing in case anyone else would like to make her or pretty blue girls in general.

Yes, shes not blue. She was mainly named after my fav weed strain. "Blue Dream" Shes more a aqua color. Her help words are msblu , aqua skin , aqua hair , aqua eyes ( if those arnt showing) , or leave them out if you want more a human toned version of her. I get good results on 0.7 personally. She tends to be nude alot, and I am working on fixing this and will share a updated version soon Ive just been super busy. Im more into making pictures then loras. Its very complicated to me lol.

( she has some issues dressing on some models for me, i havent pin pointed a fix for this, I think it has to do with the tags "aqua skin" I am planning to update soon, maybe this weekend. )

My AI art pages if youd like to follow me ( I use both AI , and Clip Studio )

https://www.deviantart.com/paintingpandaart

https://www.pixiv.net/en/users/80323848

My husbands loras here ( he's way better than me lol )

https://civitai.com/user/IndolentCat

EXAMPLE PROMPT

(vtubermodel:1.2) fullbody , 1girl, with black hair and orange eyes, smiling , (green hoodie:1), fluffy cat ears, fluffy cat tail , high quality , extra detailed , rendered, sharp face, symmetrical , facing viewer ,A-pose, (white background:1.5), looking at viewer <lora:vtubermodel_2700_lora:1.2>

neg prompt : (worst quality, low quality:1.4), border, frame, (large breasts:1.4), watermark, signature bad eyes , bad hands, asymmetrical, complex background , black background, extra fingers


this model was trained with the meinamix v9 model

trigger words: scared expression, clenched teeth, open mouth

v1.57 {based on and reg by AOM2_hardcore

It is better for you to use this model

It is born with Jack Dempa's works

}

v1.8 {has higher generalization through Nai training, and the training materials still come from jack dempa.

You will achieve better results with VAE:vae-ft-mse-840000-ema-pruned.ckpt and Model: AOM2_hardcore.

}

Please not combine this lora with realistic models.

以上

感谢ControlNet1.1的lineart功能，让这个lora得以问世

此lora仅通过AI生成的图像进行训练（制作此训练集稍微有些难度）

这次应该是独属于AI的风格了

使用时请通过color theme控制颜色（如red theme）

玩的开心！

Thanks to the lineart feature of ControlNet 1.1, this lora was made possible

This lora is trained only on AI-generated images (making this training set is a little more difficult)

This time it should be an AI-only style

To use it, please control the colours with the colour theme (e.g. red theme)

Have fun!

I tried making it.

model:abyssorangemix2_Hard

lora trained car model

v2.0 add room,bar,additional space,special effect mode.

Three modes are provided,fullbackground,simple background white background.

Support for art style switching.

Support more actions:standing,sitting,jumping,kneeling,standing on one leg,and more...

High-quality models, come and download and use them！

优质模型，快来下载使用吧！

QQ交流群：816165884

JK uniform，It is prohibited to be used for any public figure model training

JK制服，禁止用于任何公众人物模型训练

////

drak blue/white/black (long-sleeved / short-sleeved) JK_shirt

JK_style

dark blue/pink/black/red bow/long JK_tie

gray/white/black JK_skirt

dark blue/gray/black/blue JK_suit

dark blue/white/balck (long-sleeved / short-sleeved) JK_sailor

drak blue/black/gray collar

////

The ideal weight is 0.4~0.6

推荐权重0.4~0.6

Test, Work well on Anime checkpoint

his LORA mode is for Buddhists or those who are interested in Buddha statues, please do not use it in other pornographic fields, thank you。And please use it with respect. The resulting Buddha statue should try not to have incomplete limbs, or only the head of the Buddha remains. Show your whole body as much as possible. Be as dignified as possible. Amitabha.
這個LORA 模式是為了給佛教徒或是相關對於佛像有興趣的人使用，請勿用於其他色情的場域，並且請以恭敬的心來使用。產生的佛像，盡量不要出現有肢體不全，或是只剩下佛頭。盡量顯示全身。盡可能保持莊嚴。阿彌陀佛。感謝您

只是画风，无政治观点 旧存货:)

based on futaall7.0

Support me on bilibili:zwh20081的个人空间_哔哩哔哩_bilibili

My qq group:644797506 email:zwh20081@qq.com

Accept making models as a business(email me)

btw I'm 15 qwq

Droners Mouse

(recommended strength : 0.7~0.95)

The sleepy idol makes her LoRA debut! Made using Hollowstrawberry's LoRA training guide and colabs. Trained on ~100 images at around 2000 steps.

DETAILS

• This LoRA works best at weight 1 (one)

• The trigger word is kozue

  • Add white hair, green eyes to consistently get accurate hair and eyes, but this not necessary as most of the time the LoRA will do it on its own

  • Add brown dress, puffy short sleeves, buttons to get her default outfit (or something that looks very similar)

• Incredibly flexible and is compatible with many different outfits and styles

  • Compatible with other LoRAs

• Tested with AOM3 and anythingv5

• Embeddings used are EasyNegative and badhandv4 (not required)

It is a character LoRA of Senko-san from Sewayaki Kitsune no Senko-san.

While v3A (which I personally preferred before I made v4) tends to work good only with some specific models, v4 seems to be more universal. I still couldn't get any good result with 2.5D models but maybe it's just because I didn't have any experience with them.

I had to delete tags from metadata of preview images because they contain some words civitai doesn't like. Here are couple examples of images which contain PNG-info:

https://files.catbox.moe/tfu9h5.png

https://files.catbox.moe/g4uj7n.png

https://files.catbox.moe/tvxd9d.png

My favorite model for this LoRA is Anything v5 but it should work fine with other 2D models.

I tried to train new version of the LoRA using different base models, but ended with that original NAI works the best as a base model for learning.

XSarchitectural_32ModernChinesestyle

Suggestions for use:

• If you want to get the parameters, download the image and use the stable diffusion feature.

• I used realistic large models and emaVAE to help my images get good quality, you can download them from the following URL, and thank the authors for developing these files:

  • XSarchitectural-InteriorDesign-ForXSLora | Stable Diffusion Checkpoint | Civitai

  • https://civitai.com/models/4201/realistic-vision-v20

  • https://civitai.com/models/34192/ultra-sharp-high-contrast-tutorial-vaeandupscaler

If you think it's good, please remember to give me a five-star review

If you have any questions or suggestions, please contact us at xsarchitectrualstudio@gmail.com

v1 Description

Lora model of Shuten Douji from Fate Grand Order. Preview images done with weight 0.8-1.

I think this one will finally be a final version. It still have faults, but it feels more accurate than the previous one and I think I can leave it at that ^_^; I feel I bit more than I could chew cause I intended to try out different approaches here for better results, but at the same time I used bigger dataset than ever before so it was slow process. I should do something smaller next try to test more rapidly.

Also random disclaimer - keep in mind that if you copy generation data from Civitai images make sure to change lora name in tag to your lora file name. Earlier I stopped including that part in prompt because it was driving me mad to see so many get that wrong >_< it's the same for all Lora so keep it mind.

So now to prompts, they are mostly similiar to the ones before. I've ogranized them better from top to bottom for your convenience:

• Standard minimum outfit - shuten douji, headpiece, revealing clothes, bare arms, ankle ribbon, barefoot sandals. Kimono will often try to make its wy back so you can add things like purple kimono, off shoulder, open kimono to negative to ease the problem.

• Standard kimono outfit - shuten douji, headpiece, revealing clothes, purple kimono, bridal gauntlets, obi, knee pads, ankle ribbon, barefoot sandals. You can control how much kimono is open by using open kimono and off shoulder, but it's not the exact science. Also a lot depends on tags obi and revealing clothes. Just include what you want to see.

• Jiangshi outfit - shuten douji, jiangshi, qing guanmao, (ofuda:0.9), hair rings, china dress, frilled sleeves, long fingernails, sash, jingle bell, flats. Fingernails are a pain cause they just make hand problems more obvious so you may want to be careful with that. Also better do ofuda weight lower like in example cause you can get too many of them otherwise.

• Halloween caster outfit - shuten douji, forehead jewel, earrings, black choker, short twintails, dudou, rope, shimenawa, detached sleeves, shawl, star print, layered skirt, fundoshi, pelvis curtain, stirrup legwear. The prompt for rope, shimenawa may often require more weight to look right. Also it can be pretty hard to remove shawl, star print.

• Black kimono - shuten douji, headpiece, horn ornament, choker, black kimono, obi, floral print, geta, anklet. Tag for facial mark was trained too but I wouldn't bother cause it's not looking anything liek it's suppose to. Also you can try adding (checkered clothes, black scarf) but it rarely works like it's suppose to. From my experience this often needs lower Lora weight more than others.

• Maid outfit - shuten douji, maid, maid headdress, headpiece, detached collar, bowtie, puffy sleeves, black dress, wrist cuffs, maid apron, garter straps, black thighhighs, high heels. She is very likely to display upskirt if pose allows it (at least on base model I used) so you may want to add something like skirt lift, panties, holding clothes to negative.

• Travel outfit - shuten douji, headpiece, black choker, hood, black jacket, off shoulder, open jacket, pink hoodie, clothes writing, leggings, sneakers. Hoodie will often end up green because model failed to learn consistenly that green comes from inside of jacket, but I feel it still looks fine.

• Punk outfit - shuten douji, eyewear on head, headpiece, multicolored hair, collar, leather jacket, open jacket, print shirt, clothes writing, bracelet, wristband, spiked bracelet, black shorts, fishnet pantyhose, thigh strap. Details are fairly off but still looks somewhat right ^_^ eyewear on head is a hit or miss so might want to remove that if you don't want it.

v0.9 Description

Lora model of Shuten Douji from Fate Grand Order. Preview images done with weight 0.8-1.

Sadly even after I spent few days on this, it's still work in progress T_T so hopefully I'll update shortly.

I'll describe prompts better when I'm happy with it, so for now just in short:

• Standard minimum outfit - shuten douji, revealing clothes, headpiece, ankle ribbon, barefoot sandals, bare arms

• Standard kimono outfit - shuten douji, revealing clothes, headpiece, ankle ribbon, barefoot sandals, bridal gauntlets, purple kimono, obi, knee pads

• Jiangshi outfit - shuten douji, jiangshi, qing guanmao, long fingernails, china dress, hair rings, jingle bell, flats, sash, frilled sleeves, ofuda

• Halloween caster outfit - shuten douji, dudou, forehead jewel, detached sleeves, earrings, short twintails, fundoshi, shimenawa, black choker, shawl, star print, stirrup legwear, layered skirt, pelvis curtain

• Black kimono - shuten douji, black kimono, hair flower, floral print, choker, facial mark, headpiece, geta, anklet, black scarf, obi, horn ornament

• Maid outfit - shuten douji, maid, maid headdress, black thighhighs, puffy sleeves, garter straps, wrist cuffs, maid apron, high heels, bowtie, detached collar, headpiece, black dress

• Travel outfit - shuten douji, hood, black jacket, open jacket, pink hoodie, sneakers, leggings, black choker, off shoulder, clothes writing, headpiece

• Punk outfit - shuten douji, open jacket, clothes writing, bracelet, collar, wristband, leather jacket, print shirt, fishnet pantyhose, thigh strap, eyewear on head, multicolored hair, black shorts, headpiece, spiked bracelet

The last two, while not closely accurate actually work best for me.

Moka Akashiya 赤夜萌香 | Rosario + Vampire ロザリオとバンパイア

https://www.pixiv.net/users/86166786

(Red color flaming eye:1.3), Red color eye trail, flaming eye, eye trail,<lora:flamingEyeConcept_flamingEye:0.8>

https://civitai.com/models/20104/flaming-eye-concept-lora

地点：(Night:1.7), standing on the rooftop,
动作：
服饰：(green_jacket),brown skirt,black legwear, black choker, pantyhose, Deep_cleavage,
角色：<lora:MokaAkashiya-V1:0.7>,
头发：long white hair,glowing_red_eyes,
角色鎖定：
1 girl,Solo, Manga,20yo,mature female,Beautiful Finger,Beautiful long legs,Beautiful body,Beautiful Nose,Beautiful character design, perfect eyes, perfect face,
眼神表情：
looking at viewer, serious,
畫質:
official art,extremely detailed CG unity 8k wallpaper,
(masterpiece:1.0),(best_quality:1.0), ultra high res,4K,ultra-detailed,
photography, 8K, HDR, highres, absurdres:1.2, Kodak portra 400, film grain, blurry background, bokeh:1.2, lens flare, (vibrant_color:1.2)
身材:
(Beautiful,large_Breasts:1.4), (beautiful_face:1.5),(narrow_waist),
Negative prompt: EasyNegative, bad_prompt, bad-artist-anime, bad-hands-5, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry, out of focus, censorship, Missing vagina, Blurry faces, Blank faces, bad face, Ugly, extra ear, amputee, missing hands, missing arms, missing legs, Extra fingers, 6 fingers, Extra feet, Missing nipples, ghost, futanari, Extra legs, Extra hands, panties,pants, (painting by bad-artist-anime:0.9), (painting by bad-artist:0.9), text, error, blurry, jpeg artifacts, cropped, normal quality, artist name, (worst quality, low quality:1.4),
Steps: 27, Sampler: Euler a, CFG scale: 7, Seed: 3076081841, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

My first attempt to make LORA.
Control tag for this is MHXA. Tested on AOM3ab1 and Anythingv5PrtRE
Recommended power of LORA is 0.8
To invoke her second ascension outfit use MHXA, jacket, bike shorts
For first one use MHXA, red scarf, coat
Blond hair tag might be included if you get different hair color. I didn't include it in samples
Please share your images. And let me know if you want other ascension included.I am planning to add Idol version if i have time

1. 本模型是用40多张浮世绘和戏曲剧照训练得到的

   The model was trained using over 40 woodblock prints and theatrical photographs.

2. 手部还有不少问题，会争取有空的时候尽快解决(适当减少权重能好一点.)
   There are still some issues with the hands, and I will try to resolve them as soon as possible when I have free time.(Reducing the weight appropriately can alleviate some hand problems.)

3. 如果喜欢的话，请帮忙点个赞ｂ(￣▽￣)ｄ哦

   If you like it, please give it a ｂ(￣▽￣)ｄ

4. 关于炼丹欢迎交流

   Q347200557

Marlène Jobert - French actress in the 70s and author. She's Eva Green mother.

Warning : a lot of erotic content used as sources.

Please provide constructive criticism.

You can support me here to be part of the society (https://ko-fi.com/djo226). Supporters may request model.

Prompt : MJ , high quality photography, natural lighting, canon eos 5d mark ii, photorealistic, 8k, photo by Annie Leibovitz, detailed skin, ultra realistic, bokeh, sharp features, centered in frame, facing camera, symmetrical face, ideal human, 90mm lens, f16, photography, ultra details, photo <lora:MJ Cruz:0.7>

Negative prompt: (deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime, mutated hands and fingers:1.4), (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, disconnected limbs, mutation, mutated, ugly, disgusting, amputation

Steps: 80, Sampler: DPM++ 2M Karras, CFG scale: 12, Seed: 739720113, Face restoration: GFPGAN, Size: 768x512, Model hash: 81086e2b3f, Model: realisticVisionV20_v20

Face restoration is optional, may provide better result if not activated (more frekles).

Yuisis from Granblue Fantasy.

Since this is the first time I created it, there may be many unnatural parts.

Model : AnyLoRA

Prompt : gbfyuisis

V2 updated: it is more stable while the style has changed a bit. Try and compare to get your favorite!

It is an experimental result but I think it works pretty well.

Seems that the AI are, without a doubt, an Emma-chan oshi...

Please, leave your comments!

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

Demo images use Danger Zoo ~ Johnsons Negative TI Collection ~

"modelname style" is the main trigger word
additional trigger words are listed to keep you close to the trained style/concepts.

This LORA is designed for used with other djz LORA's and/or on top of Mixed pairs of djz CKPT's or a Triple merge mix including other CKPT models. This is a specific method for merged CKPT models while running a stack of these Style LORA models on top. You may still use this alone and it is recommended to include the additional trigger words in your prompts.

DJZ Season 2.5

Demo images use Danger Zoo ~ Johnsons Negative TI Collection ~

"modelname style" is the main trigger word
additional trigger words are listed to keep you close to the trained style/concepts.

This LORA is designed for used with other djz LORA's and/or on top of Mixed pairs of djz CKPT's or a Triple merge mix including other CKPT models. This is a specific method for merged CKPT models while running a stack of these Style LORA models on top. You may still use this alone and it is recommended to include the additional trigger words in your prompts.

DJZ Season 2.5

Demo images use Danger Zoo ~ Johnsons Negative TI Collection ~

"modelname style" is the main trigger word
additional trigger words are listed to keep you close to the trained style/concepts.

This LORA is designed for used with other djz LORA's and/or on top of Mixed pairs of djz CKPT's or a Triple merge mix including other CKPT models. This is a specific method for merged CKPT models while running a stack of these Style LORA models on top. You may still use this alone and it is recommended to include the additional trigger words in your prompts.

DJZ Season 2.5

Demo images use Danger Zoo ~ Johnsons Negative TI Collection ~

"modelname style" is the main trigger word
additional trigger words are listed to keep you close to the trained style/concepts.

This LORA is designed for used with other djz LORA's and/or on top of Mixed pairs of djz CKPT's or a Triple merge mix including other CKPT models. This is a specific method for merged CKPT models while running a stack of these Style LORA models on top. You may still use this alone and it is recommended to include the additional trigger words in your prompts.

DJZ Season 2.5

All images were generated using abyssorangemixs-2-Sfw

一款可以生成景观平面设计的lora模型，基于1.5版本模型训练

A LORA model that can generate landscape design layouts can be generated using a LORA model trained on version 1.5 of the model

test

My third model lora

Training model：AnythingV5V3_v5PrtRE

Sampling method：Euler a

suggest lora set 0.9-1.0 and add "Jewelry Bonney" in prompt

Have fun!

Yasaka | High School DxD

https://www.pixiv.net/users/86166786

地点：(Night:1.7), Cyberpunk_CityView, Before Window,
动作：sitting at attention,sitting on couch,
服饰：(yellow_japanese_clothes:1.3), animal ears,fox ears, tail, fox tail, multiple tails,fox girl,thick eyebrows,
角色：<lora:yasakahighschooldxd-V1:0.7>,<lora:dxdAllGirlsGasperRiasSonaYasakaVenelanaAkenoAsia_dxdv2:0.3>,Yasaka,yasaka dxd, <lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：cat ears, orange eyes, hair ornament, ribbon, hairband, yellow eyes, very long hair, blonde hair, long hair,
角色鎖定：
1 girl,Solo, Manga,20yo,mature female,Beautiful Finger,Beautiful long legs,Beautiful body,Beautiful Nose,Beautiful character design, perfect eyes, perfect face,
眼神表情：
looking at viewer,
畫質:
official art,extremely detailed CG unity 8k wallpaper, perfect lighting,Colorful, Bright_Front_face_Lighting,
(masterpiece:1.0),(best_quality:1.0), ultra high res,4K,ultra-detailed,
photography, 8K, HDR, highres, absurdres:1.2, Kodak portra 400, film grain, blurry background, bokeh:1.2, lens flare, (vibrant_color:1.2)
身材:
(Beautiful,large_Breasts:1.4), (beautiful_face:1.5),(narrow_waist),
Negative prompt: EasyNegative, bad_prompt, bad-artist-anime, bad-hands-5, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry, out of focus, censorship, Missing vagina, Blurry faces, Blank faces, bad face, Ugly, extra ear, amputee, missing hands, missing arms, missing legs, Extra fingers, 6 fingers, Extra feet, Missing nipples, ghost, futanari, Extra legs, Extra hands, panties,pants, (painting by bad-artist-anime:0.9), (painting by bad-artist:0.9), text, error, blurry, jpeg artifacts, cropped, normal quality, artist name, (worst quality, low quality:1.4),
Steps: 27, Sampler: Euler a, CFG scale: 7, Seed: 1681945853, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

test

test

test

test

基于NAI训练的星穹铁道角色——卡芙卡的模型

使用方法：

推荐权重0.7~0.9，推荐使用block weight与高清修复

核心触发词为：1girl, official, purple eyes

建议使用触发词：1girl, official, purple eyes, (empty eyes:1.2), eyewear on head, black jacket, jacket on shoulders, shirt, shorts, pantyhose, boots

translate by GPT：
Based on NAI training, the model of the Honkai Star Rail character, Kafka.

Usage:

Recommended weight is 0.7-0.9, and it is recommended to use block weight and high-definition restoration.

The core trigger words are: 1girl, official, purple eyes

It is recommended to use trigger words: 1girl, official, purple eyes, (empty eyes:1.2), eyewear on head, black jacket, jacket on shoulders, shirt, shorts, pantyhose, boots

weight 0.7 is best.

I created this LoRA in response to a request.

fiona1, spy x family, hair_over_one_eye, solo, brown gloves, blue coat, buttons, jewelry, turtleneck, earring, breasts, expressionless

I added an expressionless prompt word to all of Fiona Frost's sample images.

All characters in the example image are over 18 years old.

49 images as training set, use new trainning parameter combinations , recommended weights 0.6~0.8,no trigger, have fun!

Because of the TOS, can't upload pics with metadata

so I put the parameters here:

masterpiece ,best quality,intricate detail ,1girl,(petite), kisa, cat girl, kindergarten uniform, yellow hat, shy, blush,beautiful detailed eyes,large breasts,

Negative prompt: worst quality, low quality, normal quality, lowres, bad anatomy, bad hands, normal quality, monochrome, grayscale, jpeg artifacts, bad-hands-5

Steps: 40, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1167642073, Size: 512x768, Model hash: 812cd9f9d9, Model: anythingV3_fp16, Denoising strength: 0.55, Clip skip: 2, Hires upscale: 2, Hires upscaler: 4x-AnimeSharp, AddNet Enabled: True, AddNet Module 1: LoRA, AddNet Model 1: tsukimi-000007(f88f67239b61), AddNet Weight A 1: 1.0, AddNet Weight B 1: 1.0

请注意，例图里的所有人物都超过18岁

49张图片作为训练集，尝试了新的参数组合，推荐权重0.6~0.8,没有触发词, 玩的开心！

Trained on 52 photos, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6 on ChillOutMix. (haven't tried on other models). weight 1 is recommended for better results.
lightweight LoRA, works on any generator (easydiffusion,stablediffusion etc). share your results.

Thank you for downloading.please leave a rating and comment if you like my models. also leave a comment on what model would you like to see next. enjoy

Trigger word: yotsuba sayo, blonde hair

Outfits: sports bra, pink jacket, school uniform

Suggested weight: 0.6-0.8

*The dataset includes mainly screencap from anime and there isn't many dynamic poses from her. Please consider using strong pose prompts or lora to achieve poses you want.

Sayaka Katsuragi 葛木さやか | Jitaku Keibiin

https://www.pixiv.net/users/86166786

地点：(day:1.6),st the school, in the classroom,
动作：sitting at attention,sitting on the school chair,
服饰：red_bowtie, school_uniform, shirt,long_sleeves, pleated_skirt,,thighhighs, black_legwear,
角色： <lora:Katsuragisayaka-V1:0.7>,
头发：Black_long_hair,Black_eyes, high ponytail,
角色鎖定：
1 girl,Solo, Manga,20yo,mature female,Beautiful Finger,Beautiful long legs,Beautiful body,Beautiful Nose,Beautiful character design, perfect eyes, perfect face,
眼神表情：
looking at viewer,(Closed_mouth:1.1),(light_smile:0.5),
畫質:
official art,extremely detailed CG unity 8k wallpaper, perfect lighting,Colorful, Bright_Front_face_Lighting,
(masterpiece:1.0),(best_quality:1.0), ultra high res,4K,ultra-detailed,
photography, 8K, HDR, highres, absurdres:1.2, Kodak portra 400, film grain, blurry background, bokeh:1.2, lens flare, (vibrant_color:1.2)
身材:
(Beautiful,large_Breasts:1.4), (beautiful_face:1.5),(narrow_waist),
Negative prompt: EasyNegative, bad_prompt, bad-artist-anime, bad-hands-5, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username, blurry, out of focus, censorship, Missing vagina, Blurry faces, Blank faces, bad face, Ugly, extra ear, amputee, missing hands, missing arms, missing legs, Extra fingers, 6 fingers, Extra feet, Missing nipples, ghost, futanari, Extra legs, Extra hands, panties,pants, (painting by bad-artist-anime:0.9), (painting by bad-artist:0.9), text, error, blurry, jpeg artifacts, cropped, normal quality, artist name, (worst quality, low quality:1.4),
Steps: 27, Sampler: Euler a, CFG scale: 7, Seed: 2448932589, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

训练量少，step低，质差勿喷,图示为权重lora_koreandolllikeness:0.6,和本loar权重0.4配合使用,大模型是chilloutmix

使用的模型最好混过Pastel

Best effect:Pastel-Mix

Lora for Cornelia from FE3H. Activation tag is cornelia arnim. If you have trouble with the outfit, try dress, bare shoulders, plunging neckline, feathers.

First training attempt.

• It's definitely not perfect and prompts are slightly messy.

• Used a prompt auto-fill without really tweaking it; Eventually soon I'll go back and fix it.

• High resolution images used.

• Haven't tested on many models. (Koto, Orange and AV3/AV4 work so far)

• I personally have pretty good success using AnythingV4 with DDIM sampler at weight 0.9. Probably best around 0.6-0.9.

• Some of the preview images were accompanied by ControlNet1:1 while also using CLAMP style. No inpainting used.

• Used Hollowstrawberry's Data Set Maker and Lora Trainer on Google Collabs.

• Prompts
  I also used negative textual inversions: EasyNegative, BadHandsv5 and bad-artist.
  -------------------------------------------------------------------------------------------
  

  Some Negative Prompts that may or may not help - depending on what you want:
  
  Art nouveau
  2000s \(style\)
  White Background
  Wings
  Close Up
  (Artist name)
  Skinny
  Narrow waist
  Bottomless
  Topless
  Torch
  Border
  Cropped
  Monochrome
  2girls/3girls/multiple girls <-- (use this if you want just one character to appear)

One of my first LoRA.

How did I do it: (Russian language): https://dtf.ru/howto/1660668-obuchenie-modeli-s-pomoshchyu-lora

In the guide, I used this version of the colab: https://colab.research.google.com/github/Linaqruf/kohya-trainer/blob/0365b4045cc64a150497a742f844bd3a1bf89a8d/kohya-LoRA-dreambooth.ipynb

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。

Shauntal from Unova's Elite Four, trained on animefull

Works on 0.6-1 weight.

自训练lora，欢迎使用

Self trained Lora, welcome to use

欢迎加星球，一起分享和讨论 https://t.zsxq.com/0dnI4f5nT

也欢迎进群讨论