Lora trained on 384 images of Circe from Fate.

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:

circedef, circernd,

Trained on NAI, gen'd on AOM2 hard.

Trigger word : berlin00

Money Heist Series #1

Berlin

Thank you for visiting my post <3 Please leave review images for feedback

Lora for Luna from doaxvv. This was trained with 饭特稀_v0.8 which you can find on civit. The lora should work ok on similar models but it looks best on 饭特稀_v0.8. Trigger word is "luna".

Not trained on any particular costume. Mainly trained on her one alt hairstyle you see in my sample images.

Also highly recommend EasyNegative which my sample images are using. If you get bad looking faces at lower resolutions try inpainting only the face at higher resolution or using hires fix.

King of hypnotist in Isekai 異世界催眠王

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：white dress,bare_shoulders, cleavage, collarbone, corset, frills,
角色：<lora:Isekaihypnotist-V1:0.7>,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：brown_hair,long_hair, single braid,
角色鎖定：
1 girl, 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,(innocent_big_eyes:1.0),alluring,light smile,embarrassed, shy,
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-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: 3964229630, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

• This Lora provides adult faces for characters and blurs backgrounds it also provides better lighting. I trained with anime models so I don't know how it will perform in realistic models.

• Recommended use is around 0.45-0.7 I personally use 0.6.

• For better results "mature woman" or "mature man" is recommended

This is a LoRA extraction of my DREAD model, Version 4. For the information and training keywords, check out the details from that model:
https://civitai.com/models/14722

Animal frens in the style of Yoshizaki Mine.

About this LoRA
The updated version LoRA makes character much closer to the intended character but it's not quite there to prompt them with just their name. Considering the dataset is just one single image for each character and there's 61 of them it works better than expected though.
If you use more tags to describe the characters you can get very close or on point (see uploaded file for tags). Another option is to use it as a general style LoRA to make cute Kemono Friends knockoff girls.

How to use
Base prompt: 1girl, [character name] (see below),
Recommended strength: 0.7 to 1.0 depending on model, prompt, other LoRAs etc
If you're after a result best matching the official chibi art use a model with good medium and drawing technique tag support. Base NAI works if you have nothing else on hand.

Misc
512 resolution trained on NAI without noise offset option for best LoRA compatibility. Dataset of based on 61 images.
Tested on NAI + AOM2.

All images are raw untouched txt2img generations

Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<

Below are all characters trained in this LoRA. The list also doubles as a wildcard generator for users of the AUTOMATIC1111 release.

Note that the "\" before parenthesizes might be necessary for your GUI to know that "kemono friends" is part of the name and not a separate tag.

(1girl)(facing viewer)

(

(

{serval \(kemono friends\)|

caracal \(kemono friends\)|

common raccoon \(kemono friends\)|

fennec \(kemono friends\)|

ezo red fox \(kemono friends\)|

silver fox \(kemono friends\)|

japanese crested ibis \(kemono friends\)|

thomson's gazelle \(kemono friends\)|

white rhinoceros \(kemono friends\)|

royal penguin \(kemono friends\)|

emperor penguin \(kemono friends\)|

gentoo penguin \(kemono friends\)|

rockhopper penguin \(kemono friends\)|

humboldt penguin \(kemono friends\)|

dhole \(kemono friends\)|

meerkat \(kemono friends\)|

common dolphin \(kemono friends\)|

blackbuck \(kemono friends\)|

tasmanian devil \(kemono friends\)|

australian devil \(kemono friends\)|

koala \(kemono friends\)|

tiger \(kemono friends\)|

lion \(kemono friends\)|

grey wolf \(kemono friends\)|

japanese wolf \(kemono friends\)|

cheetah \(kemono friends\)|

scarlet ibis \(kemono friends\)|

sivatherium \(kemono friends\)|

black rhinoceros \(kemono friends\)|

arabian oryx \(kemono friends\)|

african elephant \(kemono friends\)|

indian elephant \(kemono friends\)|

leopard \(kemono friends\)|

black leopard \(kemono friends\)|

puma \(kemono friends\)|

giant panda \(kemono friends\)|

brown bear \(kemono friends\)|

okapi \(kemono friends\)|

golden snub-nosed monkey \(kemono friends\)|

bald eagle \(kemono friends\)|

northern goshawk \(kemono friends\)|

frilled lizard \(kemono friends\)|

mexico salamander \(kemono friends\)|

southern tamandua \(kemono friends\)|

polar bear \(kemono friends\)|

hippopotamus \(kemono friends\)|

bearded seal \(kemono friends\)|

japanese otter \(kemono friends\)|

small-clawed otter \(kemono friends\)|

red kangaroo \(kemono friends\)|

white tiger \(kemono friends\)|

plains zebra \(kemono friends\)|

lesser panda \(kemono friends\)|

reindeer \(kemono friends\)|

sika deer \(kemono friends\)|

moose \(kemono friends\)|

margay \(kemono friends\)|

ocelot \(kemono friends\)|

tanuki \(kemono friends\)|

northern white-faced owl \(kemono friends\)|

eurasian eagle owl \(kemono friends\)}

)

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

宝相庄严

A character LoRA for the B-Class hero Captain Mizuki from One-Punch Man. She wasn't in the original webcomic version of One-Punch Man and hasn't been in the anime yet, so there wasn't as much to train on as I'd like. Trained on Anything-v3. Usually works with a strength of 1, but it will need to be lower to avoid burning images for some models.

I haven't been able to get the bandaid on her cheek to generate correctly, it might be better not to prompt it and then inpaint it.

I also forgot to remove the tag "highres" from the dataset before training, so that might help with prompting.

LoRA trained on NAI with screencaps from the anime series Mieruko-chan, and some additional fan art. It can generate 4 characters + ghosts.

Please rate and give me some feedback if you'd like to see more of these models. And let me know if it needs improvement.

How to use:

• Trigger words for the characters: yotsuya miko, yurikawa hana, niguredou julia, and as a small bonus with much less data, yotsuya kyousuke. You can also prompt just ghosts (using dark aura and black eyes can help with the ghost effect). See image gallery for reference (generated with Hassaku model).

• Recommended 1.0 weight and your usual CFG. Reduce CFG/weight if in some models like AOM2 gets a bit blurry, and/or use hires fix.

• It is not easy to generate the different characters in the same image without some details wrong. Perhaps latent couple or inpainting would help.

• When prompting for ghosts you should also include male tags like 1boy or multiple boys. Proper prompting is required to keep ghosts and characters separate. Otherwise you can get characters as ghosts. If that's your thing, it works great :)

Characters (images below generated with AnyLora):

yotsuya miko

yurikawa hana

niguredou julia

yotsuya kyousuke

Misery Stentrem ミズリィ | Youkoso! ようこそ！

ミズリィ・ステントレーム Youkoso! Skebe Elf no Mori he.

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： a black and white outfit with gold trim around her neck,bat_wings,elbow_gloves,
角色：<lora:MiseryStentrem-V1:0.8>,long_pointy_ears,dark_elf, elf,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：blonde hair,blue eyes ,pink_lips,Bangs,(hair_Covered_one_eye:1.3),
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),alluring,light smile, embarrassed, shy,
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-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: 353477669, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

All images were generated using abyssorangemixs-2-Sfw

Note: I don't know why, but it can't generate the tail of tigress, if you put 'tail' amongst the tags it will generate one but not attached to the body.

I will try to fix that later.

lora of indie Vexroia the sun eater

This is a (beta) lora for only anthropomorphic/anthroified/pokemorph reshirams. Reasoning given at end.

This is currently an early release because I've trained several versions with several dataset changes; reshiram is a terrible subject due to all the hair and feathers and it's hard for me to find clean samples. I do have plans to grab some "lesser" images and tag those (with edits to isolate the subject/clean up background noise) and continue training later. Sadly I have schoolwork and medical business to deal with, so I cannot give an ETA.

Also, image #4 is black because I wanted to recreate a bad gen form an earlier model that looked cool and to end racism.

Use with any AnythingV3-based models and a strength of 0.8-0.9.

Tagging just "reshiram" works well, but if the chokers or feathers give you an issue then include some of these: white body, feathered body, winged arms/feathered arms, feathered wings/arm tuft, chokers, white hair, long hair, finger claws, tail/big tail. Tags with slashes should be used alternatively. The entire internet has also decided that reshiram is a milf, so males are difficult (trying to fix) but can be forced out with strong weights for male and muscular with negatives for female and female features, but it isn't consistant. I'll be working on fixing the dataset when I have time. There are no feral tags, see below.

As for the lack of regular non-bait reshiram:

I originally had a mix of feral and anthroified content in the dataset, and while it wasn't polluted by any specific images it seemed to have weird results even with an extreme anthroified bias. Forcing ((((feral)))) and negating pokemorph and anthroified did work, but the results all had four legs or turned into an anthro reshiram with a reshiram-horse hybrid (very good results but weird). Furry models were "better" but still put feral reshiram on four legs or cut it in half (???). I continued working on it and tweaking tags until I realized something: no one uploads sample images for my work, likely because it's too raunchy to post. I thought of that in the context of what's basically just an animal, and after a long 12 seconds of staring into space I wiped the feral images off of my drive and retrained without them. Instantly fixed everything and upped quality to boot. I might be a degenerate freak, but I have standards and will not create something that even tangentially might lead to beastiality/zoophilia content. Sure, it's an imaginary animal, but that doesn't change much and I don't feel comfortable with that. From here on out, any Lora's I create will only be of humans and anthros. I will commit degeneracy in the name of BIG FUNNY but I will never stoop to perversion.

Thanks for understanding.

tag:1girl,maid,black choker, hair ornament,black thighhighs, japanese clothes,detached sleeves,

This is a LoRA for generating images of Persona 5 character Okumura Haru.

Making models takes a lot of time and hard work. If you like this LoRA, please click the like button!

Recommended weights:

• Recommended weights: 0.5~0.7

  (Note: If the weight is below 0.4, the LoRA may not work effectively.)

這是一個用於生成女神異聞錄5角色奧村春的圖像的LoRA

製作模型需要花費很多的時間和精力。如果您喜歡這個LoRA，請幫忙按一個“讚”！

建議權重：

• 建議權重：0.5〜0.7

  （注意：如果權重低於0.4，則LoRA可能無法生效。）

(As Requested)

I need to search in Japanese to get better photos of her, but there aren't much besides press conference photos therefore there might be light interferences. you can try to use your own prompts to reduce the light effects.

Trained on 51 photos with ChillOutMix, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6. 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

miyamoto madoka, heroine of adult comic "Kare to Watashi to Tenchou no Shinya Kinmu", Cyclone (Izumi, Reizei).
The title means "Him, me, and the store manager's late nights"

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

To support me you can buy me a coffe at https://ko-fi.com/antonioriolo

Training settings:

33 images, 15 steps, 10 epochs, you find many epochs (with test images using plot xy script) on my Patreon while supporting me.

Recommended settings:

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

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

• CFG Scale : 5 to 9

• LORA weight for txt2img: reccomended 0.5

• Keys: wearing samba dance dress, carnival party

This model has average difficulty to be used. You just have to play with weights to get what you want. You might want to give a weight to the keyword wearing samba dress dance, 1.2 or 1.3 might be good.

Also you might want to give a weight to the dress color, since it usually draw it in yellow and green.

Example "woman, (wearing samba dress dance:1.2), (red dress:1.2), etc"

carnival keyword doesn't always work, also in this case you might want to give an higher weight to make your woman inside a carnival party, but in most of the cases i tryied it wasn't necessary. carnival or carnival party both should work.

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

Lora trained on 188 images of Deepthroat inspired by Mizumizuni

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.

Dataset was de-censored by me.

Prompts:

deepthroat / cum / hands on another's head / pov hands / head grab

To help remove it from mouth use: penis on face, penis awe,

Trained on NAI, gen'd on AOM2 hard.

Form for requests.

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

A weight of ~1.0 is recommended. Lower it to ~0.8 if you want to be more flexible with clothing. It still struggles with some colors in my experience.
Feel free to experiment with other weights and share your results if you want to!

The main prompt is "Houshou Marine"

For her outfits try:

Base Outfit:

"houshouBase, heterochromia, red eyes, yellow eyes, twintails, long hair, hair ribbon, large breasts, white gloves, frilled choker, red ascot, leotard, leotard under clothes, red jacket, cropped jacket, sleeveless jacket, black coat, off shoulder, bicorne, red skirt, miniskirt, leather belt, black thighhighs"

Gothic:

"houshouGothic, heterochromia, red eyes, yellow eyes, twintails, black ribbon, large breasts, mini top hat, hat flower, gothic lolita, short dress, red dress, frilled dress, detached sleeves, frilled sleeves, corset, bowtie, black gloves, pocket watch, white thighhighs"

Bikini:

"houshouBikini, heterochromia, red eyes, yellow eyes, ponytail, long hair, jewelry, baseball cap, sunglasses, eyewear on headwear, black jacket, open jacket, white shorts, short shorts, red bikini, string bikini, o-ring thigh strap"

Band:

"houshouBand, heterochromia, red eyes, yellow eyes, bangs, long hair, streaked hair, shako cap, two-sided jacket, off shoulder, elbow gloves, lace-trimmed leotard, bodystocking, high-waist skirt, showgirl skirt, thigh boots, large breasts"

Officer:

"houshouOfficer, heterochromia, red eyes, yellow eyes, large breasts, bangs, short hair, sleeveless shirt, collared shirt, black skirt, side slit, pantyhose, id card, fingerless gloves"

Let me know if you have any issues.

Used embeddings for sample pictures:

EasyNegative

bad_prompt_version2

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

Super Sonico すーぱーそに子

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

地点：(Night:1.5),Concert,Cyberpunk,neon lights, stage_light,
动作：standing next to a guitar,Standing on stage, holding microphone, singing , popstar,
服饰：headphones on her ears,pantyhose, purple_legwear,miniskirt,a backpack,a tshirt with a pikachu on it,
角色：<lora:supersonico-V1:0.7>, <lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：Pink_hair,long_hair,Bangs,
角色鎖定：
1 girl, 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,
(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-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: 644409663, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

I'm testing the theory that maybe the reason MidJourney's hands are so much better now is that they just took the time to specifically train a network on a high quality set of pictures of hands, and literally nobody else has actually tried. This LoRA definitely isn't at MidJourney levels yet, but I've been training it over night for several nights now and adding to the dataset where it appears deficient, and quality seems to be steadily improving. As such, I'm going to post this now so people can start using it. Consider this an early alpha -- I'll only stop updating once it stops getting better.

Example images are cherry-picked. Please don't expect this model to make all of your hand generations better. It may even make some of them worse, so you should evaluate its usefulness on a large number of images and not just one. If it works for you like it does for me, the a lot of your results should be the same or better quality (some will just be bad in different ways).

Amamitsuki Style 天 三月 - LoRA

• Good:

  • Amamitsuki style of girl face and body.

• Bad (I'll try fix it in version 2):

  • Some postures (mostly NSFW) can present problems such as multiple hands, multiple legs, etc...

  • perfer to generate picture with white background

Trigger word:amamitsuki

Recommeneded Weight: 0.6

Model: I personally like AOM3, but as you can see, it can works well on most of models. Preferable that the style of the model is close to the AmaTsuki's artstyle

GCF Scale: 7

I Used to writing a lot of prompts, but i think little also work even better.

Some Test Examples

masterpiece, best quality, ultra-detailed, illustration, portrait, best Light and Shadow quality,4k,8k,[cg],,,<lora:TabiStyle-000009:0.8:OUTALL>,tabi, (pink blue theme),,,

((very wide shot)), (from below:1.4), (full body:1.3),,,

1girl walking,

[long pink hair,beret,(black mouth mask),looking down,

black bracelet,(medium breasts),(((bralines))),((((high-waist sideboob)))), (((pink and white jirai kei))), ((gothic lolita)),lolita fashion,black frilled ribbon-trimmed long dress, white ribbon, thick thighs,((black) boots, ),((black) fishnets,),(((dynamic pose))),]

,,

<lora:AmamitsukiStyle-000006:0.7:ACT>, amamitsuki,

,,

horizon cityscape,cloudy sky, shopping district,(flower), moss,

Negative prompt: bad_prompt_version2, (((umbrella))),animal ear fluff, maid kimono,maid, kimono,red,

Steps: 18, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 4070978642, Size: 512x768, Model hash: eb4099ba9c, Model: abyssorangemix3AOM3_aom3a3, Denoising strength: 0.5, ENSD: 31337, Hires upscale: 2, Hires steps: 12, Hires upscaler: Latent, Eta: 0.667

If you'd like to support me and the LoRAs that I create consider throwing a tip my way @ https://ko-fi.com/kouyasimp

LoRA trained to replicate the meme/art trend that I'm not actually sure if it has a name for. I was originally hoping to do the whole meme but the second panel has too many variants for me to teach it to a lora so I decided to focus on the more consistent top panel. Trained on 109 images. Compatible with character and style LoRAs.

Recommended weight is 1. Recommended lowered weight if trying to make a girl or human.

Trigger word is meme but can also work without it. Can do double grabbing but difficult to produce. I'd like to make a double grabbing version in the future.

Alia Bhatt is an Indian actress who appears in Bollywood films.

Use Chilloutnight
Support + exclusive posts, LoRAs etc soon :)
https://www.buymeacoffee.com/allmywonz

A simple lora traned on 40 images of the character Morgana. Intended use on a yiffy model such as yiffymix

Spider-man into the Spiderverse Landscape Style LoRA

A LoRA Inspired by Spider-man into the Spiderverse.

• Features:

  • The graphic style of the Spider-Man parallel universe such as skewed lines, brightly coloured, etc.

  • urban at night, cityscape.

  • the artwork style.

• Good: Cityscape, some corridor, some room...

• Bad? Some times Unstable.

  • Recommended to use ControlNet for the control of the figure position and pose.

  • In some cases the wrong or rough image

Trigger word: dont need.

Recommeneded Weight: I like value around 0.6

Some Test Examples

Using abyssorangemix3, If I don't mention.

A requested pose LORA: spread armpit. I didn't remove any layers due to consistency loss but you may try this LBW if you find it hard to fit with other LORAs: 1, 0,0,0,0,0,1, 1, 1,1,1,0,0,0,0,0,0

Left and right versions are designed for latent couple+composable lora use case but later I found it extremely hard to use. The title image is the only successful one among hundreds of attempts. (Tittle image is currently flagged for review😢 so I post it on pixiv: https://www.pixiv.net/artworks/107396234 )

Suggested weight: 1

Trigger word: spread armpit

Please consider put moles in negative.

Please use lower denoising strength when applying highres fix to avoid bad fingers. I'm using 0.4 or 0.45 for these examples.

-- -- -- -- --

Request queue: airi sakura, estelle, kabedon view from side, zenos, le malin, maria naruse

Sources only from Kill Bill 1 !

( Data set from months ago so ... ) Beta! Trigger: umabird

Now taking $$$ for LoRA training for active patreons....
https://www.patreon.com/nucleardiffusion

or please consider an easy one off donation here: https://ko-fi.com/nucleardiffusion

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

Warning: First of all, do not expect this Lora to work completely well, it has given me good results and it has given me bad results, going finer it would be easier to obtain better results constantly, in any case with the freedom that this Lora gives It is normal that things do not happen as you expect, anyway I publish them in case it helps someone. Exactly as the title indicates, the parrot seeks to create more times, or more life to the drawings. I didn't see many parrots like this and it's really complex to ensure that they don't all look identical. In any case, I enjoyed the parrot and tried to play with it, without much expectation.

To access the full uncensored version of my Loras visit:
 +) https://huggingface.co/Pepomariuomddddd

The more training, the more weight the model has, I obtained the best results by indicating two instances, first the sexual one, then the subject's clothing. But who knows, you should try your luck

A LORA model trained on Oscar Lopez.

To use place the words "oscarlopez person" in the prompt.

Trained on 5620 steps.

这是一个不完美的版本，测试过各种模型有点拼运气。

我应一位本网站网友的需求制作了这个lora。

但是可收集的素材太过有限。即便我将收集的素材经过二次加工处理，但还是经常出现烂手和服饰不符合原型的情况。希望未来有机会再加改进。

测试权重在：0.6-0.8较为合适。

触发词：Chizuru Kagura

This is an imperfect build, tested various models with a bit of luck.

I made this lora at the request of a user of civitai.com.

But the material that can be collected is too limited. Even though I have processed the collected materials through secondary processing, there are still often bad hands and clothing that do not conform to the prototype. Hope to have the opportunity to improve in the future.

The test weight is more appropriate at 0.6-0.8.

Trigger word: Chizuru Kagura

Onmyoji_Shiranui, buzhihuo, 不知火

The more training, the more weight the model has, the Lora should always go well when you describe what you are looking for, if you want a blowjob, describe a blowjob, if you want the veil only, do not describe the blowjob and move away from the action of the blowjob: for example indicating to go, walk, ect

To access the full uncensored version of my Loras visit:
 +) https://huggingface.co/Pepomariuomddddd

Like all my Parrots = Loras, it is trained roughly, so it should be flexible and at the same time contain "mistakes" or "little joys"

The model not changes the expressions, anyway 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

Trigger word is swsnwtrpr. If you are getting wrong color add "white" before the trigger word. If you are getting unmasked people, add "mask" to the prompt.

There are a lot of guns in the dataset, if you dont need them add "gun" to the nagtive prompt.

For me it worked best at 0.8 weight.

Huggingface: https://huggingface.co/sheldonxxxx/MechaDream

Support me on Ko-fi to help me buy more GPU time: https://ko-fi.com/techadventurex

MechaDream

A Stable Diffusion model for Mecha

All in one model (ckpt+lora+vae in one safetensors): https://civitai.com/models/47135

Recommended Setting

Checkpoint: Counterfeit-V2.5

VAE: mecha_v2_e3-pruned.ckpt (https://huggingface.co/sheldonxxxx/MechaDream/resolve/main/mecha_v2_e3-pruned.ckpt)

Sampler: UniPC (Personal preference)

CFG Scale: 15

Negative Embeddings: EasyNegative, bad-hands-5

Lora Weight: 0.8

Hires.fix with scale >= 2 is highly recommend

My workflow

Step 1: Batch generate images

Step 2: Select the best image and inpaint any flaws

Step 3: Upscale using MultiDiffusion

🡳 Some comparison images between original image and after the above workflow 🡳

https://imgsli.com/MTcyNDM2

https://imgsli.com/MTcyNDM3

https://imgsli.com/MTcyNDM5

Roadmap

MechaDream v1 will be the base model for generic mecha image generation

New Lora models train on different anime Blu-ray will add flavors into my MechaDream

List of WIP anime:

1. Gurren Lagann (Dataset is ready, will start training when I'm less busy)

2. Code Geass

3. Macross Frontier

4. Evangelion

5. Genesis of Aquarion (Maybe)

6. Valvrave

7. Eureka Seven (Maybe)

8. Gundam

9. Armored Core

NGA荣誉版主: 喵小吉 LoRA。

Miaoxiaoji LoRA.

此 LoRA 过拟合了部分内容，可以尝试更低的权重。

This LoRA will overfit something. You can try to set too low weight.

与检查点模型 ChilloutMix 和 dalcefo_painting 工作良好。

Good work with ChilloutMix and dalcefo_painting checkpoint model.

如果脸部崩坏，尝试放大脸部。使用提示词 (detailed face, detailed eyes) 或者使用更高的分辨率.。

If the face is broken, try zooming in on it. use prompts (detailed face, detailed eyes) or use a higher resolution

使用这个lora，自动帮你设计并生成具有中心对称特征的扑克牌图像，

经过测试，此lora兼容其他角色lora，至少兼容我自己的几乎所有角色lora

强烈推荐使用Anythingv4.0，推荐Hires fix

分辨率为512×768

玩的开心！

Use this lora to automatically help you design and generate playing card images with centrosymmetric features.

This lora has been tested and is compatible with other character lora's, or at least almost all of my own character lora's.

Anything v4.0 is highly recommended and Hires fix is recommended.

Resolution is 512 x 768

Have fun!

under_table plus:

• pov: centered pov

• from side: from side, centered, you can see both sides of the desk

• from side pov: off center pov,

stealth_fellatio: add it to any of the other tags to help make other unrelated characters.

Hi all, this is a sister project of AnimeFus - Kyo Mix v0.3, which trained with anime characters from KyoAni into a single model. On the other hand, AnimeChar aims to remake the character with higher quality using Lora based on our dataset.

In Oshi Mix, we would like to introduce our new trained Lora model: Ai Hoshino (星野 アイ Hoshino Ai), a titular character in the Oshi no Ko manga series. Although the training process is a long jounery, with filtering out many unsatisfied data images, parameters, underfiited or overfitted models, we still hope this is a balanced model.

RIP Ai Q.Q

You can check character information at below link:
https://oshinoko.fandom.com/wiki/Ai_Hoshino

Please check "About this version" to know more details about the model version and included characters. The model base is Anything v4.0.

Enjoy and cheer :)

-- Please also support --
AnimeChar - Liz Bird Mix | Kasaki Nozomi (傘木 希美)

AnimeChar - Tamako Mix | Kitashirakawa Tamako (北白川 たまこ) & Asagiri Shiori (朝霧 史織)

AnimeFus - Kyo Mix v0.3

2023,04,21 v2.0

两个衣服能正常画出来了，私服那个白黑+红裙子的难画出来。

大衣：Kannagi-Yuuri-co, red coat

水手服：Kannagi-Yuuri-se, serafuku

私服：Kannagi-Yuuri-pc, red skirt

Google Translate:

The two clothes can be drawn normally, it is difficult to draw the white and black + red skirt in casual clothes.

Coat: Kannagi-Yuuri-co, red coat

serafuku: Kannagi-Yuuri-se, serafuku

Private Clothes: Kannagi-Yuuri-pc, red skirt

OLD v1.0

TAG乱套了，以后再更新吧，练了四个衣服怎么写触发词都只能画出一个校服，但衣服触发词不写只写人名到时能随机出来别的衣服......?，换衣服请用0.4及以下的强度不写提示词

Kannagi Yuuri, , serafuku, coat, civies, nightshirt.

Google Translate:

The prompt words are messed up, I will update them later when I have time, I have trained four clothes, but no matter how I write the prompt words, I can only draw one school uniform.

But if you don’t write the prompt words for clothes, just write the name of the character, and other clothes will appear randomly...?, please use an intensity below 0.4 to change clothes

RING Artstyle

• Add the prompt ring, gold, silver

• weigh : 1

• checkpoint : RevAnimated or other you like

目前还在测试阶段，欢迎评论返图，提供意见。

是一个做戒指的LoRA，用来做游戏素材不错。

可以加入一些元素，比如蛇、龙啊，可以看示例。

加入gold，silver等TAG可以改变材质颜色。

也可以加入其它物体元素，自行尝试吧。

如果你喜欢我的作品，可以点个Like，或者请我 喝杯咖啡。

爱来自瓷器。

有问题可以来这提问 https://b23.tv/MUQAevx

[机翻英语]

It is still in the testing stage, welcome to comment, return to the image, provide comments.

It is a LoRA for making rings, which is good for making game material.

You can add some elements, such as snakes, dragons and so on, you can see the example.

Add “gold”, “silver” TAG can change the color of the material.

You can also add other object elements, try it yourself.

If you like my work, you can click "Like" or buy me a coffee.

Love from China.

CAISAN Art Style

I'm making a Discord Server. Join if you want - https://discord.gg/3mGXrS87cE
Also, I need helpers to make Free Requested Loras / My Idea's Loras. Join Discord if you wanna help. this will be not paid as I don't earn anything from it or from anywhere.

Please Review and Like this Model if you appreciate it. The least you can do to keep me motivated.

Also I'll appreciate your support on my Patreon and kofi.

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

https://forms.gle/h9iuLCCnaEs6W9ZB8

trigger word : vanessa

recomm. strength : 0.88 - 1.0 (works best),

do not use below 0.8, otherwise likeness will be lost

you can add "butt" in prompt to get butt pose

if you like my stuff pls consider supporting me on Kofi

Engage Kiss

Colors - Tabi Style LoRA

Creating scenes w\ Color you want.

Trigger word: tabi

Add prompts about colors

Recommeneded Weight: I like value around 0.6

Some Test Examples And Comparisons

Using abyssorangemix3, If I don't mention.

AmamitsukiStyle & SMISV is other LoRA I trained, I may upload it later.

masterpiece, best quality, ultra-detailed, illustration, portrait, best Light and Shadow quality,4k,8k,[cg],

1girl,

dark-skinned female, white dress, white stocking, bare shoulders, white hat, (from side:1.2);

(walking),

in the middle of (((pink blue))) flower field,

cityscape, <lora:TabiStyle:0.9>,tabi,<lora:SMISV-000012:0.5>,ArtOfSMISV,

Negative prompt: bad_prompt_version2,

Steps: 18, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 1923956653, Size: 768x1152, Model hash: 5493a0ec49, Model: abyssorangemix3AOM3_aom3a1b, Denoising strength: 0.5, ENSD: 31337, ControlNet-0 Enabled: True, ControlNet-0 Module: none, ControlNet-0 Model: control_openpose [b46e25f5], ControlNet-0 Weight: 0.6, ControlNet-0 Guidance Start: 0, ControlNet-0 Guidance End: 0.3, Hires upscale: 1.5, Hires steps: 12, Hires upscaler: Latent, Eta: 0.667

Try different colors:

masterpiece, best quality, ultra-detailed, illustration, portrait, best Light and Shadow quality,4k,8k,[cg],,,

<lora:TabiStyle-000012:0.6>, tabi, (pink theme),,,

10mm lens f4, landscape, (very wide shot), (from below:1.1), (full body:1.4),,,

horizon structure, (outdoors,), cloudy sky, flower field, grasslands, lake,,,

( 1girl walking),((dynamic pose)),,,

Negative prompt: bad_prompt_version2,

Steps: 18, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 1923956653, Size: 768x1152, Model hash: 5493a0ec49, Model: abyssorangemix3AOM3_aom3a1b, Denoising strength: 0.5, ENSD: 31337, Hires upscale: 2, Hires steps: 12, Hires upscaler: Latent, Eta: 0.667

On Different Models:

masterpiece, best quality, ultra-detailed, illustration, portrait, best Light and Shadow quality,4k,8k,[cg],,,

<lora:TabiStyle:0.6>,tabi, (pink theme),,,landscape, (very wide shot),harbor, sea,

( 1girl walking),((dynamic pose)),,,

Negative prompt: bad_prompt_version2,

Steps: 15, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 911702946, Size: 512x768,ENSD: 31337, Eta: 0.667

Recreate style of Margot Robbie Harley Quinn costume.

Flexible enought to use it in combo with others lora, check the chibi version :D

In my examples i used "Lora Block Weight" with this preset: MIO:1,0,0,0,0,0,0,0,1,1,1,1,1,0,0,0,0

( Data set from months ago so ... ) Beta! Trigger: nklul

Now taking $$$ for LoRA training for active patreons....
https://www.patreon.com/nucleardiffusion

or please consider an easy one off donation here: https://ko-fi.com/nucleardiffusion

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

other characters from anime:

Astarotte Ygvar

Asuha Touhara

Naoya Touhara

Ingrid Sorveig Sorgrims

Judith Snorrevik

Griselda Reginhard

Elfreda Mirjasdottir

Cuthfleda

Olaf Frismar

Elika Drakul Draupnils

Lucca Austri

Unnbjorg Signar

Mistrune Asgrim

Mercelida Ygvar

Ursula Sumarlidi

Isold

Sigurd Sveinnsson Svarthaed

Isadora Finnsdottir

Porhelga Svarthaed

Backgrounds

Dragon

Fii

Enja Níuhali

Teacher

Atli Reginhard

Miina-sensei

Naoya Touhara_trap

Lama

Works fine with all sd 1.5 models, train on anything V5 [Prt-RE]

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

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

Donate

boosty

It makes things decorated with colorful stained glass.

Perpetual coffee request: I signed up for Buy-Me-a-Coffee and Ko-fi, mainly because I thought it would be hilarious if someone as artistically untalented as myself received a coin for something creative. It's already happened, and it made my day. AI made the impossible possible :D. If you'd like to add to the amusement, please consider buying me more coffee through the following links:

https://www.buymeacoffee.com/konyconiN

https://ko-fi.com/konyconi7992

If you like it, follow me on Twitter as I post my work there.

https://twitter.com/AIEroticGirls

I recommend using weights (0.6 - 0.8) for better results.

It is also recommended to mix it with "lora:Japanese-doll-likeness."

Resolution 512x512

※Please use the product in compliance with the laws of each country.

IF YOU ARE NEW TO STABLE DIFFUSION OR YOU HAVE PROBLEMS WITH MY MODELS, GO HERE: Guide

IF YOUR PICTURES HAVE PROBLEMS, CHECK MY FAQ SECTION: FAQs

RIGHT NOW IMAGES METADATA ARE MISSING AS I'M USING COMFYUI TO SIMPLIFY THE IMAGE GENERATION. I'M WORKING ON A WAY TO FIX THIS USING A CUSTOM NODE.

What's this LoRA

This is a LoRA trained on low poly images made with Bing Image Creator. My models are not trained to achieve the exact style of their source, but to change the general vibe of the image. They are also fine-tuned on anime models, so they may not work perfectly with realistic ones.

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

• https://civitai.com/models/24149

• https://civitai.com/models/26332

• https://civitai.com/models/28322

• https://civitai.com/models/32022

• https://civitai.com/models/33707

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

The LoRA works also on NSFW (check out the examples).

Guides

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

• Beginner's Guide

• Prompting

• My Workflow

• Inpainting

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

Support

Do you want to get access to my LoRAs before everyone, and contribute to my next models? Check out my supporter tiers on Deviant Art:

If you want to support my work for free, you can also check out my music/art here:

• Bandcamp

• Instagram

• DeviantArt

other characters from anime:

Astarotte Ygvar

Asuha Touhara

Naoya Touhara

Ingrid Sorveig Sorgrims

Judith Snorrevik

Griselda Reginhard

Elfreda Mirjasdottir

Cuthfleda

Olaf Frismar

Elika Drakul Draupnils

Lucca Austri

Unnbjorg Signar

Mistrune Asgrim

Mercelida Ygvar

Ursula Sumarlidi

Isold

Sigurd Sveinnsson Svarthaed

Isadora Finnsdottir

Porhelga Svarthaed

Backgrounds

Dragon

Fii

Enja Níuhali

Teacher

Atli Reginhard

Miina-sensei

Naoya Touhara_trap

Lama

Works fine with all sd 1.5 models. Trained on anything V5 [Prt-RE]

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

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

Donate

boosty

Link (Blue Tunic - The Legend of Zelda: Breath of the Wild) LoRA

Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕

Oddly enough, there was no Blue Tunic Link on Civitai, so I made it. Should work fine on AnyLoRA, Anime Pastel Dream and NeverEnding Dream at around weight 1. Trigger word is link but using the legend of zelda, the legend of zelda: breath of the wild and blue tunic might help too. As usual, I didn't overload the trigger word, so you should still use descriptive words. This will allow you to use this "link toolset" to make new characters based on Link, or also have Link wear different stuff or improve your other Link lora/embeddings. This will also allow you to use it better with Latent Couple, as it will conflict less with other characters.

You can also make a female version of link like I did in the last examples.

How to use LoRA's in auto1111:

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

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

• Select your LoRA like in this video

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

sumo-e ( a kind of ukiyo-e)

About Honami > Here

Recommended

Scale 5-15
Steps : 35-50
Weight: 0.4-1

Warning
Half of examples are soft NSFW

Test Model

Anything V 4.5

My FB Page : fb.com/aikengkador

tag:1girl,sarong,revealing clothes

因为某种需求顺带炼制着这个lora，这个角色是《雄兵连/超神学院》里的莫甘娜，当然我更喜欢叫她凉冰，毕竟那句“我是恶魔之王，随心所欲.....”

太中二了....

但我还是很喜欢这个角色，极具个人魅力，当然也非常的好看。

导演你快复活她好吗？

别逼我跪下来求你好吗？

所以如果你也喜欢这个角色，欢迎你使用这个lora，权重推荐0.7左右

tag是liangbing，

底模我使用的事rev或者tmsyMIDREALV

当然你也可以尝试其他的，真人风格

（训练素材会有一点模糊，随意我推荐你们可以加一些提高清晰度的lora或者加大尺寸或者tag，这是我采用的解决方案，仅供参考）

最后，这个lora只做同人作品分享，请勿使用与任何商业用途。

Finally, this lora is only for sharing of fan works, please do not use it for any commercial purpose.

Portraits of Asian business men

Dwight Kurt Schrute III is a fictional character on The Office (U.S.) and is portrayed by American actor Rainn Wilson.

Prompt : Dwight, 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:Dwight: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

Recommend using DreamShaper, with weight of 0.7.

The trigger keyword is bettyboop girl

for example:

a bettyboop girl ,holding a wand

Trigger word is ck_underwear.

This LoRA is incomplete.
LoRA's pose and body are limited.

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.

Introduction

A LoRA model trained with ~700 images by Gaston18. The trigger tag is ‘gaston18’ (trigger tag is optional actually). The 784mb VAEs (NovelAI, Anythingv3, Orangemix, Counterfeit) are recommended. 0.6 ~ 0.8 weights are good. The preview images are generated using 1.6 version (original) YesMix.

Cautions (Must Read)

• Some preview images are generated using 'Hires Fix' and upscaled using 'SD Upscale'. Thus, if you wanna generate the preview images, pls download the original preview images and analyze them using 'PNG INFO' in webui to get all generating information.

• The maximum of generating resolution should not exceed 1024 theoretically, otherwise extra/deformed head/limbs will be waiting for you :). Try to use 'hires fix' to prevent this problem and help to generate higher-resolution images.

• For 784mb VAEs (NovelAI, Anythingv3, Orangemix, Counterfeit) users, make sure add '--no-half-vae' in command line of webui to prevent generating black images. See voldy's log for the details of edit method. For high-performance GPU, make sure 'medvram' and 'lowvram' mode are disabled.

Re-upload because previous LoRA has some trouble.
Just a LoRA for pose a girl holding phone taking picture with mirror.
Tested, work well on anime model like CounterfeitV25
Some of useful keyword: {sitting | lying | kneeling}, {from behind | looking at viewer}, {pulling clothes, topless, bottomless, lingerie}, holding phone

Recmmend weight; 0.7, you can try others for your purpose, combine with other LoRA for better results

Feel free to ask me for training your LoRA (style, person, character). Support me for anothers LoRA https://www.buymeacoffee.com/dwang

Form for requests.

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

A weight of ~1.0 is recommended.
Feel free to experiment with other weights and share your results if you want to!

The main prompt is "Itsumi Erika"

Use "bangs, medium hair" for her hair style.

For her outfit you can try the following prompt:
"Itsumi Erika, bangs, kuromorimine military uniform, black jacket, red shirt, red skirt, hat"

Let me know if you have any issues.

Used embeddings for sample pictures:

EasyNegative

bad_prompt_version2

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

Lora model customization: click here

Animation series: Click here

Version Description

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

The model is made based on ch modal。

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

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

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

1girl, blue eyes,white hair, best quality, photo, 4k,looking at viewer,upper body,(wear t-shirt),

Role of Arknights,and which is the NEW COSTUME,

Weight 0.8~1.0 ,and you can ADJUST prompt to reductive character

TRIGGER:

1.BASE: SKADITEA;

2.Costume: black gloves, black shirt, black thighhighs, shorts, open green jacket

One of Skadi's casual outfits.

MARTHE Sports Line New Arrivals/The Next Afternoon Tea.

Iizunamaru Megumu from touhou 18 Unconnected Marketeers.

No big change from the last version, but the training style bit changed, and it looks much better.

Tsukasa and Chimata's LoRA has also been updated to v2.0. Please check it if you're using it!

※ Trigger words:

Iizunamaru Megumu

blue tokin hat

white circle scarf

blue dress

golden pauldron

black cape

pom pom (clothes)

purple tengu-geta

※ Most of her images will appear with round ears. Use the trigger word "pointy ears" if you prefer this way.

※ You may add "holding tripod" to a negative, if it appeared unexpectedly.

It can also be used as a trigger, but won't work that well.

(I excluded the training data with tripod, or removed tripod as much as I can.)

※ Recommended weight: 0.75~0.85, CFG Scale: 5~7

(Not always the best, better try adjusting from this and find the best condition for your model.)

※ Clip skip: 2

※ Trained with animefull

What about some disturbing things ?

2.1 RZ - MACABRE -  An horror Lora model

I'm heavily using the forAnalogPortrait negative embeddings, it prunes almost all digital stems.

Note : demo pictures was generated from Providence - Anchor. Some prompts are really messy, it comes from my own trained gpt-2 model which is still W.I.P

Troubleshoot

If you are getting the error "failed to match keys when loading Lora" it means you are either on the native Lora integration of Automatic1111 which may not support LoCon, please install the https://github.com/kohya-ss/sd-webui-additional-networks extension to solve the issue.

Happy tuning everyone !

romerorz.art - Twitter - DeviantArt - Gumroad

I create my loras for free if you want to support me feel free to buy me a coffe

I am excited to share with you a model of a character from Project Moon game franchise Angela! Trained on Nai very flexible and listen to prompts well. The activation tag is angela \(project moon\).

Tenkyuu Chimata from touhou 18 Unconnected Marketeers.

Re-trained with better training data, and simpler Danbooru-style tags.

It will work quite differently from the last version, with more stabilized details.

Tsukasa and Megumu's LoRA has also been updated to v2.0. Please check it if you're using it!

※ Trigger words (Changed!):

Tenkyuu Chimata

multicolored hairband

multicolored dress

zipper lines (clothes)

cape

cloud print

brown bag belt

pink boots

※ Recommended weight: 0.75~0.85, CFG Scale: 5~7

(Not always the best, better try adjusting from this and find the best condition for your model.)

※ Clip skip: 2

※ Trained with animefull

Kudamaki Tsukasa from touhou 18 Unconnected Marketeers.

Re-trained with better training data, and simpler Danbooru-style tags.

It will work quite differently from the last version, with more stabilized details and her tail.

Megumu and Chimata's lora has also been updated to v2.0, please check it if you're using it!

※ Trigger words (Changed!):

Kudamaki Tsukasa

fox ears

fox tail

hair between eyes

white romper

green ribbon

white socks

※ No necessary Negatives, but adding "long hair" to negative can make her hair more stable.

※ You may also add "holding test tube" to a negative, if cylinder in her hand appeared unexpectedly.

It can also be used as a trigger, but won't work that well.

※ Recommended weight: 0.75~0.85, CFG Scale: 5~7

(Not always the best, better try adjusting from this and find the best condition for your model.)

※ Clip skip: 2

※ Trained with animefull

weight 0.9 is best.

I created this LORA as requested.

hancock1, 1girl, boa hancock, large breasts, long hair, epaulettes, cape, crop top, long sleeves, side slit

<without coat>
hancock1, 1girl, boa hancock, large breasts, long hair, crop top, long sleeves, side slit

I found that realistic models can hardly generate good faces from side. So I trained a simple Lora to make it happen.

You may add control prompt such as "facing to the side", "facing up", "looking at the side", "looking down"...etc. to have better control of the direction.

"Looking at viewer" is not recommended, which may generate unexpected result.

nsfw is also not supported in this version, I may update to v2.0 to include it (when I got time).

经过测试模型的适应力还算不错，能适用各种服装

通常服装触发词为nnk，和服服装触发词为zhengyuennk

例:

通常服装：

1girl, animal_ears, animal_hands, black_legwear, black_skirt, bow, closed_mouth, elf, frills, hairband, hand_on_hip, looking_at_viewer, paw_gloves, pointy_ears, puffy_short_sleeves, puffy_sleeves, shirt, skirt, solo, suspender_skirt, suspenders, tail, very_long_hair, white_background, white_shirt, nnk

和服服装：

1girl, architecture, blue_sky, blue_umbrella, bridge, building, castle, city, closed_mouth, cloud, cloudy_sky, day, east_asian_architecture, flower, holding, holding_umbrella, long_sleeves, looking_at_viewer, oil-paper_umbrella, outdoors, pagoda, parasol, pink_umbrella, pointy_ears, purple_umbrella, red_umbrella, sky, smile, solo, town, umbrella, very_long_hair, zhengyuennk

服装列表：

1.飞鱼服。2.敦煌飞天服

注意：我的模型经过 locon 训练，所以需要安装LoCon 支持。https://github.com/KohakuBlueleaf/a1111-sd-webui-locon

PS：有好点子可以留言，如果你下载了Lora，希望能留言告诉我您的使用感觉和困惑，好让我能够学习。（求5星好评，让更多人能够用到。）

PS: Leave a comment if you have any good ideas. If you have downloaded Lora, I hope you can leave a comment telling me how you feel about using it and how confused you are so that I can learn.

我的Ko-fi网址，里面有一些我做的图片，欢迎来看看。

My Ko-fi site, which has some pictures that I made, feel free to come and have a look.

https://ko-fi.com/asseer

【Note: 支持性捐助仅适用于经济独立的人群，任何人都可以下载图片但没有那么高清。Supportive donations are only available to those who are financially independent. All can download the images but not in as much high resolution.】

【PS：感谢每一个帮我买可乐的朋友，投喂1欧元=7元钱，可以一个月内下载所有我做的2K+的高清图片。PS: Thanks to everyone who helped me buy a coke by throwing in 1 euro = 7 dollars to download all the 2K+ HD images I made for a month.】

声明：本Lora中已经通过手动抹除的方式去除人脸（我一张张擦的），所以不含有任何人脸图像（也是总是会生成衣服的原因），AI生成的人脸皆为使用者自带模组或和我无关的软件生成。视频中的衣着来源，是通过百度搜索获得的包括但不限于街拍，去除商标和可识别特征的只剩分块布料的宣传剧照，国家典藏馆公开展出的文物照片等。同时呼吁，不要用AI做违法背德的事情，社区环境靠大家。

Disclaimer: The faces in this Lora have been removed by manual erasure (I erased them one by one), so there are no images of any faces (which is why clothes are always generated), and the AI-generated faces are all generated by the user's own modules or by software unrelated to mine. The sources of the clothing in the video are obtained through Baidu searches including but not limited to street photography, promotional stills with only pieces of fabric removed from the logo and identifiable features, photos of artefacts on public display at the National Collection, etc. I also urge that AI not be used to do illegal and unethical things, the community environment depends on everyone.

我的图例，均由下列tags生成/my pictures are all from prompts below.

正向词：Prompts

{best quality}, wonderful detail, solo, handsome/beautiful face, female/male focus, photorealistic, (solo),

1boy/1girl, young teenager, long hair, accurate eyes and mouth,

<任何外貌模组/any appearance model>,

<lora:Feiyufu-飞鱼服:0.8>, feiyufu, chinese clothes, dress embroidered with animals and [flowers]（OR any things you want/(衣服上绣什么可以加，但这个版本没那么精准）, black and blue theme, holding sword（OR any action/任何动作）,

china architecture, landscape, <任何可以添加风景的模组/landscape things>.

负向词：Negative Prompts

NSFW, lowres,stamp,bad anatomy,bad hands, red, error, missing fingers,extra digit, fewer digits, cropped, worstquality, low quality, normal quality, jpegartifacts, signature, watermark, username,blurry,bad feet, faceless, no character, blue eyeball, frame, easynegative,text, multiple boys, （自己调试/Debug by yourself）

Euler a 采样方式=Euler a

Steps=30 采样次数=30

Face Fix 面部修复

CFG scale=7 相关性=7

Model= Anything4.5/Stable Diffusion v1.5

Fubuki (fakemonkey0224) style LoRA

Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕

Super cute artstyle I found recently while looking at TLoZ artworks. should work well on AnyLoRA ckpt at weight 1. Use trigger fubuki \(fakemonkey0224\) for better accuracy.

How to use LoRA's in auto1111:

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

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

• Select your LoRA like in this video

• Make sure to change the weight (by default it's :1 which is usually too high)

―――【鮮やかなり天元の花。
　　　その剣、無空の高みに届く。】

All Outfit LoRA of Musashi from FGO, works around 0.8-1.0 weight

FGO武藏的全外观模型，推荐权重0.8-1

Honestly, this LoRA was not as difficult to train as I thought.

比我想象的要好训练很多

-

Supports 10 sets of Official Outfits in total, 3 sets of Saber Class, 3 sets of swimsuit Berserker class, also 4 sets of Craft Essences.

支持共计10套官方外观，剑阶3套+狂阶3套+礼装外观4套

-

Saber1= Saber1, detached sleeves, Default Kimono

Saber2=Saber2, Advanced Clothing, handguard

Saber3=Saber3, Final Costume

Berserker1=Berserker1, One-Piece Competition Swimsuit

Berserker2=Berserker2, FlagPattern Clothing, belt

Berserker3=Beserker3, Fairy Style Outfit, white sleeves, swimsuit, detached collar

Travel Portrait: Miyamoto Musashi英灵旅装=TravelingOutfit, HikingClothin, bare arms, jean

Storms and Waves疾风怒涛=StormsandWaves, Lace Bikini

See Emptiness as the Path以空为道=EmptinessPath, NewYearKimono

Sightseeing Portrait: Miyamoto Musashi英灵纪行=TravelJournal, Musashi‘s Casual Outfit, striped skirt

-

Feel free to post any idea or character model request in the Discussion

有模型的需求可以B站私信或者在这里留言给我

id：不吃青蛙的阿龙

有空就会做

If you like my work, feel free to support me on patreon.com/RiuKiMK1

or RiuKi_MK1 is Creating AI artworks, LoRa models (buymeacoffee.com)

My Second Lora: Mattew Mercer

TL;DR:

1. This is my second attempt at training a LoRA - Would love CONSTRUCTIVE feedback

2. Trained using rMadArt (SD1.5) (v4.0 | 2BD42F315F specifically)

3. 13 High Res images used, hand captioned

4. 1,300 total Steps

5. Should I make more LoRAs?

6. There will like likely be a second version of this with more training data provided

Settings:

• CFG: Seems pretty flexible, but thoroughly tested on 5 - 9 CFG

• Weight: LoRA weight works fine at 1.0, but as per most LoRA, 0.7 gives you the best results

• Model Checkpoints: Although trained on rMadArt, I've tested with a number of other models and it seems to play very nicely

• Trigger Word: No trigger word needed, but if you have issues, use 'Matt Mercer', especially at the front.

Usage:

• Non-commercial

• Don't be a dick with your renders

• The intention of this LoRA if for creating tasteful art and fan art

• Any request by Matt Mercer to remove this will be honoured immediately

Contact/Questions:

Send through to contact<dot>hotdog800<at>gmail<dot>com

First: A quick public service announcement from a random person on the internet you don't know: smoking is bad for your health. Please don't smoke.

Now onto the description: I use this at 1.0 weight, though it will blend better with other loras at 0.9 or even 0.8. However, the lower the weight gets, the more it seems to do weird things. All images were tagged with smoking so it should be a reliable trigger. The holding_cigarette tag will try to push it more towards images with a cigarette in their hand instead of their mouth. Conversely with a cigarette in their mouth should push it more towards, you guessed it, a cigarette in their mouth. In both cases it will sometimes completely ignore that instruction for reasons I don't know. Adding smoke will make it more likely to generate a plume of smoke somewhere, and if you're lucky sometimes that smoke is even coming from their mouth or from the cigarette. Use lots of negative prompts to help weed out some of the common problems like mutated hands, blurry images, etc.

Making a Lora for smoking is surprisingly difficult. This is as close as I could get after several attempts, and multiple days of GPU training time. It's definitely not close to perfect, but it can generate a usable picture about one in five times. I noticed someone else on civitai tried to make a model for smoking and ran into similar issues. I think the problem is that there are a huge variety of positions, grip styles, cigarette styles/brands, etc, that all qualify as "smoking". I kept adding images hoping that the model might eventually figure this all out, and ended up with over 130 images. The data set includes people of various genders, races, and ages, though admittedly it could still use more diversity. I had previously also included paintings and cgi art, but just a few of those skewed the whole Lora to producing only paintings, so the current data set only includes photos of real humans. The data set has cigarettes in the mouth, in the hand, holding it like a pen, holding it between fingers, holding it in their mouth with their hand, lighting the cigarette with a lighter, lighting it with a match, etc. I also used a handful of regularization images to refine what a cigarette even looked like in the first place (though it could definitely use more of them, that fixed some issues where a cigarette looks like a magic wand). After all that, it still has a hard time consistently producing a simple "woman smoking" image without a ton of positive and negative tags. I'm kind of exhausted with training this model and moving on to something else, but I may come back to it at some point. If someone else picks up the baton here, I might suggest trying to narrow down the definition of "smoking". Pick a specific grip style for example or if you want a cigarette in the subject's mouth then only train on images where the cigarette is in their mouth.

The pros:

• It can usually make an object that looks like a cigarette.

• Sometimes that cigarette is actually in their hands or mouth as you would expect.

• It should (hopefully) work with a wide variety of people of different ages, genders, and races

• It is often pretty good at making smoke, and that smoke is sometimes even near the cigarette

The cons:

• Sometimes the cigarette is being held in an awkward position in the subject's hand.

• Will often deform the hands and seems to make the model "forget" what hands should actually look like.

• Will sometimes turn one or more fingers into a cigarette.

• Will sometimes remove whatever object is being held in the other hand (like a cell phone) (reduce the weight to 0.9 or 0.8 to help fix that)

• Will sometimes make multiple cigarettes (one in mouth, one in hand)

• A cigarette in the mouth may be at a super weird angle, or not even align with the lips

• The cigarette is often not lit or shows no ash

• The smoke it makes is sometimes in the wrong place (smoke near their hand when the cigarette is in their mouth, or just in some random other place)

Combining it with a TI that can fix hands might help with some of the hands mutating issues.

Recommend using DreamShaper, with weight of 0.7.

This model has two characters and supports the presence of two people in one screen.

(The effect of the group photo is not particularly good)

We need to define the roles in this format: "a popeye man" or "an olive girl"

for example:

a popeye man, on the beach

This model is about supporting personages of ASTAROTTE NO OMOCHA! anime.

Characters in versions.

Small datasets WARNING.

other characters from anime:

Astarotte Ygvar

Asuha Touhara

Naoya Touhara

Ingrid Sorveig Sorgrims

Judith Snorrevik

Griselda Reginhard

Elfreda Mirjasdottir

Cuthfleda

Olaf Frismar

Elika Drakul Draupnils

Lucca Austri

Unnbjorg Signar

Mistrune Asgrim

Mercelida Ygvar

Ursula Sumarlidi

Isold

Sigurd Sveinnsson Svarthaed

Isadora Finnsdottir

Porhelga Svarthaed

Backgrounds

Dragon

Fii

Enja Níuhali

Teacher

Atli Reginhard

Miina-sensei

Naoya Touhara_trap

Lama

Works fine with all sd 1.5 models. Trained on anything V5 [Prt-RE]

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

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

Donate

boosty

Catria from Fire Emblem, trained on animefull.

Tested at 0.4-1 weight, easter catria works with 1, but with the other ones starting low seems to produce better results. I had a small amount of images to work with for catria_echoes and catria_white so I apologize if those give consistency issues.

Highly recommend to add "(hair ornament, hair flower:1.2)" to your negative prompt if not generating easter catria as it has a tendency to add the flowers of her costume to other versions.

catria_nm for the first Fire Emblem Heroes Catria look

catria_easter for her bunny alt from Fire Emblem Heroes

catria_echoes for her Shadows of Valentia look

catria_white for the Fire Emblem Heroes alt costume that gives her white thighhighs

Check example images for the relevant tags that help get specific costumes more consistently.

Akagi Shun style LoRA

Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕

A very good artist I saw recently. It works great with my AnyLoRA model at weight 1. I suggest you add akagi shun and watercolor to your prompts.

How to use LoRA's in auto1111:

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

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

• Select your LoRA like in this video

• Make sure to change the weight (by default it's :1 which is usually too high)

Shion | Reincarnated As A Slime

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

地点：(Night:1.7), high-rise_modern_apartment, Bedroom,
动作：sitting at attention,sitting on couch,
服饰：Deep_cleavage, Puple_long_sleeves_jacket,collarbone, (Puple_Pencil_skirt),
角色：<lora:Shion slime-V1:0.7>,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORAv2:0.2>, oda Non,
头发：Purple_Hair,ponytail,(Black_single_horn:1.1),
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),(expressionless:1.3),
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-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: 2141671991, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

He is a renowned surgical expert in the whole universe, known for his superb scalpel methods

Poor AMD, when will you come forward?

就要色色！

tag:1girl,heterochromia,tattoo,japanese clothes,katana,blue eyes,red eyes

In the future, if someone asks whether it is convenient to disassemble the clothing, let's do v2 again...

以后如果对于服装是否方便拆解，有人提要求了，再做v2吧……

Try do do a Hell look like Lora model, and see how it goes

3-Stripe Slippers Mix

Suggest to use over 0.7 weights, with a trigger word "3stripe slipper".

This LORA was trained on close-up feet images wearing the famous 3-stripe slippers, so it may not look well on full body or images other than close-up feet or leg focused images.

Works okay on SD 1.5 and some mixed models (e.g. HenmixReal_V20, etc.) but you have to roll your dice a lot to get a desirable image.

Good luck and I'll come back with V2 in the future, if this what people want.

https://linktr.ee/promqueen1

Suggeded rating: 0.6-0.9

Trigger word: iwamine chinari, black long hair, fake rabbit ears

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

Training settings:

35 images, 25 steps, 10 epochs, you find many epochs (with test images using plot xy script) on my Patreon while supporting me.

Recommended settings:

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

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

• CFG Scale : 5 to 9

• LORA weight for txt2img: reccomended 0.4-0.7

• Keys: wearing ballet dress, wearing tutu, in theater, in a studio, ballet pose (doesn't always work), ballet corp (very hard to make it work)

This model is quite easy to use, just start your prompt with "a ballet dancer wearing ballet dress", add if you want "wearing tutu" and change the color as you want like "a ballet dancer wearing red ballet dress, wearing yellow tutu".

Keywords "in theater" and "in a studio" might give good results considering its a ballet dress and thats the proper environment for it.

regarding ballet corp keyword, i didn't manage to make it work even if many images of ballet corps dancing were used to train the model, you might give it a try and achieve what i didn't.

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

This is a character model from the anime Aoki Hagane no Arpeggio: Ars Nova (Arpeggio of Blue Steel).

Tested in various chekpoint, concepts... poses and conditions. Well, you know what I mean ;-)

To start, just type in Lora and past in prompt.

Examples of art are made on the main known anime checkpoints. This art i maked in OrangeCoCoa. Without VAE.

If you have "NansException error occurs: a tensor with all NaNs was created in VAE" then add the --no-half-vae argument to webui-user.bat in the set COMMANDLINE_ARGS= line

This is my first public LoRA, it specializes in generating images of dark elves that are both visually stunning and highly attractive. This LoRA is capable of producing lifelike and detailed depictions of dark elves in a variety of different settings and poses. Whether you are looking to create erotic art, fantasy illustrations, or game design assets, the LoRA model is the perfect tool for generating high-quality images of sexy dark elves that are sure to capture the imagination and inspire the senses.

这是我非常喜欢的一个韩国女团明星，为此分享一个lora给大家下载。

Treyni | Reincarnated As A Slime

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

地点：a green background with overgrown leaves,green_background, green_theme,
动作：
服饰：green and white dress,Deep_cleavage,
角色：<lora:treyni_slime-V1:0.7>,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：green hair ,blue eyes,bangs,

角色鎖定：
1 girl, Manga,25yo,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, (innocent_big_eyes:1.0),(expressionless:1.3),
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-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: 2585202722, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original

It's Katya, Bloodsucker of the Night.

There were few pictures for training, so it is overfitting.

The weight of the LoRA may be set as 0.5-0.7 for a qualified output.

My bilibili space: space.bilibili.com/84123368

My qq group: 901256738

(As Requested)

Quite hard to populate her photos . I hope it doesn't turn out so bad.

Trained on 51 photos with ChillOutMix, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6. 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

Kamio Misuzu from Air

Useful prompts are "air school dress uniform" or "plaid pink camisole"

The LoRA model dedicated to producing beautiful penises

V3.0

• Recommended

• Use 768*768 images to train the model

V2.0

• Trying to increase the diversity of the model

• Use 512*512 images to train the model

V1.0

• Use 512*512 images to train the model

Orphen from the 90's series.
Can also generate the modern version.

Prompts

Negative prompts: (If need be)

White background
simple background
painting_(/medium)
watercolor_(/medium)
blonde hair
lens flare
shirtless
2000's style
1girl
multiple girls
floating_(object) - removes part of headband

this 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

Hello, this is my first model that I create for the character u-110 from the game azur lane, it is an experimental model and do not expect a good image quality but you can experiment with it

What other character do you want from the lora? I'll be experimenting to see how my learning progress goes

I recommend a value of 0.6 to 0.7 or however you can use it so that the character comes out as well as putting

in anime style recomend value of 0.5 and 0.4

in positive

(u-110 \(azur lane\), blue eyes, silver hair,)

in negatives

(ship,submarine)

if you want me to have the default clothes:

(swimsuit, white legwear, gloves, cape,hood),

to reinforce the character's appearance

If you help me generate more high-quality images, it is a great help to further improve the model with better backgrounds and poses.

Albis | Reincarnated As A Slime

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：white and purple dress,kimono,
角色：<lora:Alvis_slime-V1:0.7>,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：long hair,black_and_blonde_hair, hair_between_eyes, multicolored_hair,
角色鎖定：
1 girl, 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-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: 2761323210, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: Latent

2023,04,21 v2.0

三个衣服能正常画出来了，第四个那个红色的大礼服画不好，不知道写什么描述词能辅助AI画出来

泳衣：Alice-sw, swimsuit

护士：Alice-nu, nurse

女仆：Alice-en, enmaided, black thighhighs

私服：Alice-pc, (我不知道怎么写能画出来，辅助不明白，就当他没有吧)

Google Translate:

The three clothes can be drawn normally, The fourth red dress is not well drawn. I don’t know what descriptors to write to help AI draw it.

Swimsuit: Alice-sw, swimsuit

Nurse: Alice-nu, nurse

Maid: Alice-en, enmaided

Private Clothes：Alice-pc, (Think of prompt words yourself, this is not easy to draw)

OLD v1.0

默认动作叉腰（气急败坏）。三件衣服：女仆，泳装，护士。

护士服装忘了修改tag文本了，直接是AI自己检测的，写bikini就行，还是能凑合画出来的。

Google Translate:

The default action is akimbo. Three clothes: maid, swimsuit, nurse.

I forgot to modify the tag text for the nurse’s costume. It is directly detected by the AI itself. Just write bikini, and it can still be drawn barely.

This is a parody of Kaori Yuki style Lora, most of the pictures are from her work < Angel Sanctuary > ArtBook.

From my results, individuals prefer weights of 0.6 ~ 0.8

There are still some shortcomings and later changes will be made

Hope you like it, welcome to leave a message

Lora จักรยานเสือหมอบ สำหรับ ใช้เพื่อการศึกษา

Training on model: realisticVisionV13_v13

also work well on manga or 2d 2.5d model

Izis | Reincarnated As A Slime

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：glasses,red dress,black-framed_eyewear,Black_bra, pantyhose, feather_boa, fur-trimmed_coat, fur-trimmed_jacket, fur_coat, fur_collar, fur_trim,
角色：<lora:Izis_slime-V1:0.7>, earrings, <lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：double_bun,green_eyes,Black_hair,
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),(expressionless:1.3),
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-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: 3392007991, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

Different tags have different effects, with a weight of approximately 0.8

大概有四种模式,权重约0.7，依旧是副产物，银色紧身衣难度太大，可以说就是个人物lora

There are approximately four modes.Still a byproduct, making silver tight fitting clothes is too difficult.It can be said that it's just a character Lora

It is not perfect. I will improve it in the future.

Do not create or share any NSFW image from my LoRA

-Trained on Deliberate v2

Frey | Reincarnated As A Slime

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

地点：(Night:1.7), Cyberpunk_CityView, Before Window,
动作：sitting at attention,sitting on couch,cleavage,
服饰：angel_wings,asymmetrical_wings, bare_shoulders,red outfit,puffy_sleeves, red_leotard, feathered_wings, fur_collar, fur_trim,
角色：<lora:FreySlime-V1:0.7>,angel,<lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：white hair,green_eyes,(Bare_forehead:1.2),(center_parted_hair:1.2), Bangs,
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),(expressionless:1.3),
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),

Élisabeth Borne is the Prime Minister of France, known for her attentive listening and her attention to the French people (sarcasm).

Please provide constructive criticism.

You can support me here to be able to improve my models (https://ko-fi.com/djo226). Supporters may request model.

Prompt : EB , 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:EB:1>

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

白泽，狻猊，

注意：我的模型经过 locon 训练，所以需要安装LoCon 支持

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

我的Ko-fi网址，里面有一些我做的图片，欢迎来看看。

My Ko-fi site, which has some pictures that I made, feel free to come and have a look.

https://ko-fi.com/asseer

【Note: 支持性捐助仅适用于经济独立的人群，任何人都可以下载图片但没有那么高清。Supportive donations are only available to those who are financially independent. All can download the images but not in as much high resolution.】

【PS：感谢每一个帮我买可乐的朋友，投喂1欧元=7元钱，可以一个月内下载所有我做的2K+的高清图片。PS: Thanks to everyone who helped me buy a coke by throwing in 1 euro = 7 dollars to download all the 2K+ HD images I made for a month.】

To find out the meaning of a sentence, copy the text below into the Google Translate website and output it from the automatic detection into the language you need.

Yuuki Nobuteru (1990) style LoRA

Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕

Another great author from the 90s. As always works great with my AnyLoRA model at weight 1. Increase accuracy with yuuki nobuteru, 1990s \(style\).

How to use LoRA's in auto1111:

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

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

• Select your LoRA like in this video

• Make sure to change the weight (by default it's :1 which is usually too high)

需要自己调整tattoo,anklet权重来还原大王花腿上的花纹和腿环

tattoo一般在0.3~0.6

anklet一般在0.1~0.4，如果发现两只脚都有环就降低点试试，但是不一定能成功……

Here's another one of love live's VA. Personally I like her ponytail hairstyle the best.

Trained on 50 photos with ChillOutMix, work best on DPM++ SDE Karras sampling steps 20-30 CFG Scale 6. 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

It's so small, but I still prefer without 😜

This is V01 version of updated Lora that you can see in my Patreon.
Preview of V02, with Full LoRA you gain consistency, smaller swimsuits, and big variation of poses.

To celebrate 69 patreons and 120 people on the discord server I announce the first prompt contest (on discord server) worth a LoRA exclusively yours!!!

LoRA prompt weight: 0.6 to 0.8
Model used for training: Uber Realistic Porn Merge 1.3.

Recomendations:

• CFG ~9

• Steps >=28

• DPM Family or DDIM

Prompts that can improve your results: flat_chest, tiny_micro_bikini...

Tendencies of this LoRA (and resolution):

• This model was made for slim girls, the results with curvy girls aren't good.

• It's common the metallic effect on bikinis, you can put "matte", "velvet" or other finish on pos prompt.

• Sometimes nipples can pop out of bikini, you can put "nipples" on neg.

• Necklaces and navel piercings can appear, I liked them, but you can put "necklace" and "navel piercing" on neg if you don't want this.

See my other LoRAs:

• Puffy Nipples

• Perfect Perky Breasts

• Tight Bikini Top

• Small Breasts [Lite]

• Big Natural Breasts Type01 [Lite]

• Big Natural Breasts Type02 [Lite]

• Ghost Nipples [Lite]

• Underwater Naked Shot

• Fake Breasts [Lite]

• Macromastia [Lite]

• Big Ass Slim Waist (PAWG ver.)

🐔

My very first LORA ever. It's probably not very good. If anyone knows how to make it more versatile and higher quality let me know. It works okay at 760x760 with pretty much any 1.5 based model.

Trypophobia is a strong emotional or visceral reaction to patterns of clustered small holes or bumps that can cause fear, anxiety, disgust or discomfort in some individuals.

Perpetual coffee request: I signed up for Buy-Me-a-Coffee and Ko-fi, mainly because I thought it would be hilarious if someone as artistically untalented as myself received a coin for something creative. It's already happened, and it made my day. AI made the impossible possible :D. If you'd like to add to the amusement, please consider buying me more coffee through the following links:

https://www.buymeacoffee.com/konyconiN

https://ko-fi.com/konyconi7992

A Character Lora for Ichiro Yamada (山田一郎) 【Hypnosis mic】

Weight:0.6～0.8

Hires. fix

trigger words: yamada ichiro\(hypnosis mic\)

Sub words:black hair, red eyes, heterochromia, green eyes, mole, mole under eye

※If there are many moles, adjust to (mole, mole under eye: 0.6)

Normal costume(19 years old) words: hood, white hoodie, open clothes, multicolored jacket, red jacket, hood down, belt, ring

※Headphones will appear around neck without being instructed. . . perhaps

The Dirty Dawg(17 years old) words: hair between eyes, red hoodie, black shirt, bracelet, piercing, gakuran, the dirty dawg\(hypnosis mic\)

※It shouldn't appear even if you don't write the headphone spell negative

Such a microphone and speaker appearance words: microphone, cable, speaker, Mc.B.B speaker\(hypnosis mic\), Mc.B.B microphone\(hypnosis mic\)

1) Please use this Lora Data at your own risk. I can not be held liable or responsible for any damages caused by using this data.

2) This Lora Data may not be used for any works which contain negative content toward the original product, individual, &/or group.

3) Do not use this Lora to create pornographic and/or violent materials.(There is no problem for personal enjoyment. Please refrain from posting pornographic and/or violent depictions using this Lora on SNS or other public places.)

I am using google translate.Sorry for the details in Japanese

※以下日本語で失礼します。

運要素が絡みますがぽい衣装を出現させることは確認しています。

【注意】

共通呪文 : yamada ichiro\(hypnosis mic\)

認識率を高める呪文：black hair, red eyes, heterochromia, green eyes, mole, mole under eye

※ほくろが多く出現する場合は呪文自体を消すか(mole, mole under eye: 0.6)など弱めて下さい。

デフォ衣装： hood, white hoodie, open clothes, multicolored jacket, red jacket, hood down, belt, ring

※ヘッドフォンは呪文を書かなくても出現するはずです。むしろ書くと首に掛けながら耳にも装着と色々事故ります。

The Dirty Dawg容姿：hair between eyes, red hoodie, black shirt, bracelet, piercing, gakuran, the dirty dawg\(hypnosis mic\)

※こちらもヘッドフォンをネガティブに書かなくても出現しないはずです。

Mc.B.Bマイクとスピーカー出現：microphone, cable, speaker, Mc.B.B speaker\(hypnosis mic\), Mc.B.B microphone\(hypnosis mic\)

（マイクのB.B部分やスピーカー柄など細部まで学習できていません。あくまでも、ぽいものとしてご使用下さい…）

1) このLora データは、自己責任で使用してください。 このデータを使用して生じたいかなる損害についても、私は責任を負いません。

2）本Loraデータは、原作者、個人、団体等に対する否定的な内容を含む作品への使用はできません。

3) このLoraを使用してポルノや暴力的な素材を作成しないでください。(個人で楽しむ分には問題ありません。このLoraを使用してのポルノや暴力的描写をSNSなど公の場に触れるのはお控え下さい。)

https://civitai.com/models/11339/aaa

this model + kl-f8-anime2 vae或者pastel-waifu-diffusion.vae

Although I also tested the others，like REFSLAVE V1,V2.....，

The weight is 0.8 to 1, and the following is also possible

C的例tag我没认真填，可以去我的P保存原图，放到SD里查看信息https://www.pixiv.net/users/23034129

This lora trained on dataset from 2 random pictures generated by other ai. Did it for the purpose of learning and arguing, because why not.
Work on realistic vision 2.0.

<lora_Odessition:(0.7-1.2)>

braces

my first attempt at an appearance lora; it works, ish. It heavily impacts style, but it works.

This is trained at 768, I might retrain at 1024; some of the detail from the braces was lost.

weight = 0.7, enjoy

The material of this model comes from the Internet, and this person does not exist in reality.

It is recommended to use the weight of lora: 0.6~0.7

model:chilloutmix
vae:840000

This is my Nude Female Portrait / Cropped Torso LoRA. This LoRA is trained on NovelAI, so it will not do particularly well with realistic models. It was also trained with a heavily NSFW dataset.

The strength is good from 0.6 to 0.8, but I use 0.75 strength. All images are generated are generated with Hires Fix and the Orangemix VAE. All pictures were made with the blends available on my profile.

The keyword to this LoRA is 'cropped torso'. Tags like 'simple background', 'white background', and 'upper body' also help.

Additionally, I now have a discord. You should join.

天凉多喝防冻液 (tianliang duohe fangdongye)
Style based on this artist a friend of mine really likes. It's biased towards younger/petite women but it also works quite well with adult women. Not really tested on men, but younger boys should be fine.

Tested on a variety of models (AOM3, Anything v3, NAI, mixprov4, etc) but seems to work best on FuzzyAlmond (hash: 0e57438eae). Enjoy!

BETA MODEL

This model is kinda hard to make because its hard to find good quality image of the twins.

i will improve it later :DDDD

this is my worst model so far 💀💀💀💀

please use : https://huggingface.co/datasets/gsdf/EasyNegative
and use higher fix on.

Weights for lora: 0.6-1 please try what works best for you :DDD

any comment, suggestions and tips are appreciated :D

This model was made as a test, so it may not be that good.

Put (service cap, flight cap) in negative prompt if you do not want the cap.

If you have any ideas for LoRA's let me know.

All images were made with High-Res-Fix and 4x-UltraSharp, some were made with Dynamic Thresholding.

A request from a Kofi donator
m1919a4 from Girls' Frontline

Keywords:

m1919a4
m1919a4Jacketless
m1919a4Vampire

Recommend 0.8-1 strength

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

Optimal values:0.5-0.8

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! :)

群号753748318，有兴趣来玩ww

权重约0.65，勉强能用，利用ai产出这个素材难度太高了

The weight is about 0.7, which is barely usable. It is too difficult to use AI to produce this material

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

No.2xx 大概会是我个人喜欢的明星或者博主。

No.4xx 是一些概念。

lora强度推荐0.6-0.8。任何模型均可出图，词条看示意图，主要是furry一个词，可以加入no humans等词增强，定义皮色实例：pink hair或pink fox，负面提示词我用了EasyNegative 文件：EasyNegative | Stable Diffusion TextualInversion | Civitai

素材来自我去年的dreambooth模型一系列产出，几乎是最早的furry模型，今天炼制为lora，兼容任何sd1.5模型。（此dreambooth模型曾在十二月于国内论坛发布，c站不再上传）：

https://www.pixiv.net/artworks/104110325

https://www.pixiv.net/artworks/105564472。

请及时下载保存，模型随时可能被删除。

你的每一个五星都是我更新的动力。

Every 5-star review of yours motivates me to update.

公共频道：t.me/dawn66666666

I've cobbled together a sergal LoRA in the hopes of seeing more art from other sergal enjoyers. These cheese dragons deserve some love!

Use weights for LoRA from 0.5 to 0.7 or it will overcook the image.

It was trained on the YiffyMix v2 model and performs best with it since it already has a good affinity for sergals. Other models I've tried were not as good.

It should be pretty versatile for both females and males. The main tag is 'sergal' but there is also a strong affinity for 'messy hair', 'fluffy', 'open mouth' and 'tongue out'.

You can add 'flat chested' for more classic sergal female gens.

The main problem, besides occasional extra digits on paws, is tails. They are long and curly, and SD very often makes double tails or even circle tails. You can try using remedies like 'multiple tails' and 'badly drawn tail' in the negatives but it's mostly a placebo.

You ask about, khm, ‘clitoral hood’? Ah, I see you're a person of culture as well. The data and tagging are there, but I did not manage to get anything that is not extremely cursed. So, sorry, maybe next time.

This LoRa is trained based on anything 4.5. If you want to join a community of AI drawing, and to learn more about LoRa training tricks, you can scan the QR code below to join our WeCaht group chat.

这个LoRa模型是基于anything 4.5 训练的。 如果你想加入ai绘画团体，并且对LoRa训练小技巧感兴趣，你可以扫描下方的二维码加入我们的“C站炼丹术分享讨论群”，欢迎大家加入并且交流讨论！！！

捏Ta QQ频道同步欢迎你的加入（复制到qq频道内搜索）

t0o1vs01m0

This LoRa is trained based on anything 4.5. If you want to join a community of AI drawing, and to learn more about LoRa training tricks, you can scan the QR code below to join our WeCaht group chat.

这个LoRa模型是基于anything 4.5 训练的。 如果你想加入ai绘画团体，并且对LoRa训练小技巧感兴趣，你可以扫描下方的二维码加入我们的“C站炼丹术分享讨论群”，欢迎大家加入并且交流讨论！！！

捏Ta QQ频道同步欢迎你的加入（复制到qq频道内搜索）

t0o1vs01m0

148 images

Actually not an empty lora

Was trained on a bunch of shitty images but came out pretty well, and a Wendigo is one of those creatures where the more faults in it the creepier it looks imo.

works good at 0.8, didnt test with any other loras.

if you want to shoot me a tip you can do it here : https://ko-fi.com/toadviner

Beta version! of Nick Valentine from Fallout 4

Escalayer | Beat Angel Escalayer 超昂天使

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

地点：(Night:1.7), stadium, star_\(sky\), starry_sky, street, tokyo_\(city\), tower, town,sky, skyline, skyscraper,river, road, rooftop,
动作：sitting at attention,sitting on the school_chair,
服饰： elbow_gloves,puffy_sleeves,sailor outfit,
角色：<lora:Escalayer-V1:0.7>,magical_girl,<lora:aiueOkaStyle_v10:0.1>, <lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：Pink_hair,blue eyes,hair_ribbon,Bangs,
角色鎖定：
1 girl, 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, (LIGHT_SMILE:0.6),
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-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: 2544922755, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

A model I decided to make because I feel like the models uploaded here already are trained on old photos of her, I've attempted to train this on newer photos. I'm also practicing making loras so this is by no means supposed to be a professional upload or an upload where it looks like I know what I'm doing.

Trained with the keyword "nomobil" because I would assume training on SD1.5 using "Billie Eilish" in the prompt would effect results but I haven't tested.

Trained prompt structure example:

nomobil with blonde shoulder length hair and bangs with blue eyes and brown eyebrows, serious facial expression, lip gloss, gold necklace with a charm on the middle, pearl earing, blurry black background

Post your results! As I haven't tested too much. Thanks!

16 images

1 watamelon

アレンビーの普段着再現版です。

LoRA多重使用にはあまり向いていません。

ファイティングスーツ版は下記をご利用ください。

https://civitai.com/models/46354/allenbyg-gundam

My First Lora: Tiana Upcheva

TL;DR:

1. This is my first attempt at training a LoRA - Would love CONSTRUCTIVE feedback

2. Trained using rMadArt (SD1.5) (v4.0 | 2BD42F315F specifically)

3. 34 High Res images used, hand captioned

4. 3,400 total Steps

5. Should I make more LoRAs?

Settings:

• CFG: Seems pretty flexible, but thoroughly tested on 5 - 9 CFG

• Weight: LoRA weight works fine at 1.0, but as per most LoRA, 0.7 gives you the best results

• Model Checkpoints: Although trained on rMadArt, I've tested with a number of other models and it seems to play very nicely

• Trigger Word: No trigger word needed, but if you have issues, use 'Tiana Upcheva', especially at the front.

Usage:

• Non-commercial

• Don't be a dick with your renders

• The intention of this LoRA if for creating tasteful art and fan art

• Any request by Tiana Upcheva to remove this will be honoured immediately

Contact/Questions:

Send through to contact<dot>hotdog800<at>gmail<dot>com

The LoRA is intended to generate pony play/human pony BDSM concept.

trigger word: pony play

Add cart may create a cart, but it fail often.

For v2. Suggest to use weighting about 0.6 to 0.9

Notice the suit of pony play is highly naked, I don't suppose it can generate not naked people.

Nanosuit similar to what you'd see in CRYSIS or other future games / art.

Playing around with the weights will get you different creative results. Weight 0.5-0.6 is a decent sweet spot for posing and still having a lot of armor. But in the end, depends on your checkpoint.

Toy Around and Have fun.

Gonna be honest, first lora. I have no idea what I'm doing I just tossed it in and hit start. Basically a buddy of mine is obsessed with Essex memes so I asked him to give me his folder and I just made a lora of everything in it. It can make some funny memes with some creative tagging. Works best on anythingv3 or v5 but there are others it can potentially work on I think.

Also use the chibi tag with the lora at weight 0.5, I probably could have trained it better to make it more flexible but it does its job of making memes and shitposts of chibi essex so I decided to post it up here.

If you're interested you can hit up my anime Discord server, I have my bot making stable diffusion art free for all to use.

My Discord Server

Trigger word: king bradley

It can draw two types of cloth.

To draw uniform specific, add military uniform, blue coat

To draw battle suit specific, add black shirt, suspenders

If the eyepatch not draw, add: eyepatch. (But might draw two...)

XSarchitectural-40PopmartMechaGirl

Suggestions for use:

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

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

Tachibana | Sakusei Byoutou 搾精病棟

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

地点：(day:1.7), infirmary, tokyo, Beside window, Cityview,
动作：sitting at attention,sitting on the school_chair,
服饰：nurse_cap, Nurse_uniform,short_sleeves,white_shirt, wing_collar,collared_shirt, white pencil skirt,
角色：<lora:TachibanaSakuseiByoutou-V1:0.7>,<lora:saiminSeishidou:0.1>,nurse,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：Blue_hair,Bangs,short_hair,
角色鎖定：
1 girl, 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, Angry,
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-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: 98391826, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

A WIP lora for the zabrak race from Star Wars. Definitely not perfect yet. Higher weights are still needed to get close.

tag:1girl,ainu clothes,bird

Based on everyone's opinions, I will strengthen the head features, hairstyle and hair accessories, etc., and make the clothes easier to replace

综合各位的意见，我会加固头部特征，发型发饰等，而让服装更好更换

推荐底模：realistic V20

推荐尺寸：512X512

推荐放大算法：ESRGAN 4x

本模型仅供学习研究，严禁用于商业用途。

Ayane Shirakawa 白河彩音 | Overflow

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

地点：(day:1.7), School,classroom, tokyo, Beside window, Cityview,
动作：sitting at attention,sitting on the school_chair,
服饰：blue_sailor_collar,white_shirt,blue_bowtie,sailor_collar, school_uniform, serafuku,long_sleeves, pleated skirt,
角色：<lora:AyaneShirakawa-V1:0.7>,<lora:saiminSeishidou:0.1>,
头发：brown_Long_hair,brown_eyes,bangs,
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),
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-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: 1418264184, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

Lola Zieta

use weight 0.6-0.8

This is my Kardia of Rhodes / Knight-Girl LoRA. This LoRA is trained on NovelAI, so it will not do particularly well with realistic models. It was also trained with a heavily NSFW dataset. This dataset was also heavily geared towards the character Kardia of Rhodes.

The strength is good from 0.6 to 0.8, but I use 0.75 strength. All images are generated are generated with Hires Fix and the Orangemix VAE. All pictures were made with the blends available on my profile.

The keyword to this LoRA is "knightgirl". Additionally, you can have things such as 'helm', 'helmet', 'armor', 'breastplate', etc. Any tag involving armor will do the trick. If you tag without nudity, you'll probably get non-nude results, but it's pretty geared towards partial nudity. Still more clothed than the average MMO woman outfit though.

Additionally, I now have a discord server linked on my profile. I'll offer advice for LoRA training and model merging there. Also a place to show off what you make.

XSarchitectural-39ComicPackDesign

An attempt to...

https://www.buymeacoffee.com/kittykangz

ChilloutNight

vae-ft-mse-840000-ema-pruned

Lora trained on 95 images of Black Dragon Iris

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:

iris1

Trained on NAI, gen'd on AOM2 hard.

Kotone Shirakawa 白河琴音 | Overflow

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

地点：(day:1.7), School,classroom, tokyo, Beside window, Cityview,
动作：sitting at attention,sitting on the school_chair,
服饰：blue_sailor_collar,white_shirt,blue_bowtie,sailor_collar, school_uniform, serafuku,long_sleeves, pleated skirt,
角色：<lora:KotoneShirakawa-V1:0.7>,<lora:saiminSeishidou:0.1>,
头发：Black_short_hair,brown_eyes,bangs,
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),
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-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: 171256076, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

Sara Vickers, first try.

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

If you like my LoRAs please consider to support me by buying me a coffee here

And if you want to see my work diary and notice post of next LoRAs, click here

The commission is open! You can request LoRA here

Model : chilloutmix_NiPrunedFp32Fix

Vae : vae-ft-mse-840000-ema-pruned

Please do not post NSFW or sexual pictures.

Noumi Kudryavka from Little Busters!

Prompt in "white kud cape" and "white kud beret" if you want her cape and hat. And don't forget the fang!

Futaki Kanata from Little Busters!

Best used around 0.9 weight

Sakura Akiyama 秋山桜 | Kyouhaku 脅迫

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：Pink_sweater, cleavage, collarbone,
角色：<lora:SakuraAkiyama-V1:0.7>,<lora:aiueOkaStyle_v10:0.1>, <lora:saiminSeishidou:0.1>,<lora:odaNonStyleLORA_v2:0.2>, oda Non,
头发：Green_hair,long_hair, blue eyes,wavy_hair,bangs,
角色鎖定：
1 girl, 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, (innocent_big_eyes:1.0),(expressionless:1.3),
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-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: 2837882386, Size: 512x768, Model hash: 3e9211917c, Model: yesmix_v16Original, Denoising strength: 0.5, Hires upscale: 2, Hires steps: 18, Hires upscaler: R-ESRGAN 4x+ Anime6B

This LoRA allows you to generate characters in the style of a Fate/Grand Order game sprite.

The LoRA was tested on various models, and the results were generally pretty unsatisfactory when used alone. I recommand using my other LoRA to replicate Ufotable artstyle, and, if you're using Automatic1111 webui, to turn on Highres. Fix ! With the Ufotable Lora and the Highres Fix, the generated images are much better than using the FGO Sprite LoRA alone. You can see the difference with the images of this post.

While using this LoRA, I recommand adding the fgo sprite tag to your prompt, to ensure to have only FGO Sprite style images.

The LoRA seems to give better results while using AnythingV5, but models like AnythingV4.5, AnythingV3 or NAI also work. It's up to you, to use the model you like the most.

The LoRA weight should be between 0.8 and 1.0, and I recommand using a 9:16 ratio, or something resembling it, since using a 1:1 ratio to generate an image let pretty much every model add some (really cool) effects in the background, or to simply generate multiple times the same character.

My 2nd LoRA, changed the training data so it became a bit big .... nevertheless!

Discover the mesmerizing world of werewolves through creating your own stunning images that capture their primal and mysterious essence. These images will transport you to a world where the moon shines bright and the creatures of the night roam free, evoking a sense of wonder and fascination that is both eerie and alluring.

Trigger word: werewolf
Sampling Method: Euler and Euler a

Han GiSol

"A Goddess from Korea that is well-known for her marvelous appearance and wonderful body posture. She is a model and dedicates her fans with her modeling photos posted on her social media accounts." I need the number for her publicist.

This is a LoRA. If you'd rather have a textual inversion of Han, please go HERE.

Like my work? Buy me a beer!

Another addition in Granblue Fantasy collection. One of the least popular characters — The Devil companion Fraux is here.

I want to thank @Hollowstrawberry for the collab he made. And also @rockerBOO, @ChameleonAI, @zykurv, @IndolentCat for their tips and tricks! Check their creations as well!

The reviews and feedback will be very valuable for me.

The main trigger word:

• gbfraux

To recreate her original outfit the next words in prompt will be useful:

• robe or hooded robe, off shoulder for showing bare shoulders, leotard (however you can skip that if you want, most of the time it will still generate something close to her original leotard-like suit), gold trim for adding additional details, pendant choker and wide sleeves.

• some models may sometimes fail to generate her ears, in this case I recommend adding animal ears to prompt to guarantee that, but most of the time it should work without it.

• another words that you can experiment with are detached pants, short shorts, detached sleeves. i did not have much luck with those unfortunately.

• i highly recommend adding simple background to negative prompt

Weights:

This LoRA was trained on AnyLora from @Lykon, so sweet spot for it is 0.7, however depending on the model you using you might want to adjust that. Especially if you will be trying to use it with realistic models.

I'm not sure what name LoRA will get on downloading, so please make sure you using the proper name in prompt of the file that you downloaded if you going to copy prompts from previews.

Negative embeddings:

Two negative embeddings were used in preview prompts:

• AuroraNegative, you can find download link for it in description on this page: https://civitai.com/models/40199/aurora

• badhandv4: https://civitai.com/models/16993/badhandv4-animeillustdiffusion

Use a level between 0.8 and 1, which is a safe range.

You could make adjustments, just don't set it too low that it generates a random Filipino woman, and remember to include the trigger word.

I hope you get the same results in your testing.

Yoon Young-Chae

Yoon is a beautiful internet celebrity of some kind, not to be confused with actress Yoon-Chae-Young.

This is a LoRA. If you'd rather have a textual inversion of Yoon, please go HERE.

Like my work? Buy me a beer!

Arknights outfit-texas

• 欢迎加入AI绘画模型博物馆QQ群

• 1群：791222249

• 2群：599962550

That fellatio something something from kurotama.
Model thinks the censor on the dick are fingers sometimes...well, damn.

There is a list of top tags on the "About this version" >>>>>

Osiris D3 2001

Add the epitome of technical skate shoe design to your images: The Osiris D3 2001. Loved and even fetishized by many, hated by others. It's a chunky and loud sneaker. Either just add the LoRA to your prompt or use osirisd3 as trigger word for a stronger effect.

Baggy pants

2001 was also the year of baggy jeans, preferably sagged low and made from silvery shiny fabric. This LoRA can be used to end the skinny jeans era enshrined in the Stable Diffusion models. Use the trigger words baggy jeans or baggy pants to get those nice wide and loose pants. Add silver for a more shiny effect.

Infos

The LoRA was trained on SD 1.5 as base model, but works well on either realistic or anime models and can also be used very well for NSFW creations. Recommended strength is 1.

IF YOU ARE NEW TO STABLE DIFFUSION OR YOU HAVE PROBLEMS WITH MY MODELS, GO HERE: Guide

IF YOUR PICTURES HAVE PROBLEMS, CHECK MY FAQ SECTION: FAQs

RIGHT NOW IMAGES METADATA ARE MISSING AS I'M USING COMFYUI TO SIMPLIFY THE IMAGE GENERATION. I'M TRYING TO FIND A WAY TO FIX THIS, IF SOMEONE KNOWS HOW TO DO IT, I'D LIKE TO KNOW.

What's this LoRA

This is a LoRA trained on blacklight photos made with Bing Image Creator. My models are not trained to achieve the exact style of their source, but to change the general vibe of the image. They are also fine-tuned on anime models, so they may not work perfectly with realistic ones.

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

• https://civitai.com/models/24149

• https://civitai.com/models/26332

• https://civitai.com/models/28322

• https://civitai.com/models/32022

• https://civitai.com/models/33707

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

The LoRA works also on NSFW (check out the examples).

Guides

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

• Beginner's Guide

• Prompting

• My Workflow

• Inpainting

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

Support

Do you want to get access to my LoRAs before everyone, and contribute to my next models? Check out my supporter tiers on Deviant Art:

If you want to support my work for free, you can also check out my music/art here:

• Bandcamp

• Instagram

• DeviantArt

If you like my LoRAs please consider to support me by buying me a coffee here

And if you want to see my work diary and notice post of next LoRAs, click here

The commission is open! You can request LoRA here

Model : chilloutmix_NiPrunedFp32Fix

Vae : vae-ft-mse-840000-ema-pruned

Please do not post NSFW or sexual pictures.

Kronya from Fire Emblem: Three Houses. Trained on NAI.

Tested at 0.7-0.9 strength.

Trigger word is "Kronya".

• Add "cutout leotard, thigh straps, detached sleeves, thighhigh boots, fingerless gloves" for her typical look.

• "Tails" adds her tails, but can be somewhat inconsistent.

• "Facial mark" adds the mark under her eye, but often adds too many marks; inpainting a mark is usually easier.

• Some models really can't handle grey skin well; add "grey skin" for consistency if you run into problems.

Tauriel, added into the Hobbit for a bit of dwarve lurve action. Trained at 768*768. 1 epoch, 20 images, can't remeber steps but think 2600? sweet spot seems to be around 0.8, seens to work well with most samplers and seem much happier with res around 768. I think Taur1al is a prompt word, but haven't yet tried the lora without so??? Legolas next maybe?

LoRA trained on art by artist Yabby (얍비): twitter pixiv patreon
Trigger: Yabby
Weight: 0.7 worked well for me, 1.0 has much artifacting
Negative prompt recommended: tentacles, tail
Model: any NAI-like model should work


Detail:
First time training LoRA. I did not manually adjust the automatic tags from AI tagger, so there is some fighting you will need to do with the prompting, see my sample images for some guidance.
I may make an update with better training data if I find the time.
The artist primarily draws big boobs and Sona/Ahri so there is inherit bias towards that but it does seem to be able to steer away still.

Training information:
checkpoint: NAI, res: 512x512, dataset: 145 images, repeats: 15, epoch: 10, batch 3
duration: ~2hrs on RTX 2070 Super

This time I used more trigger words to make the lora more accurate and flexible.

Trigger words: ph Pardofelis, ph Pardofelisds, ph Pardofelishd, 1girl, cat ears, solo, blue eyes, brown hair, breasts, bikini top only, braid, navel, black bikini, boots, black gloves, short hair, fingerless gloves, black skirt, cat tail, bangs, asymmetrical gloves, bare shoulders, headband, green eyes, tattoo, midriff, cross-laced footwear, elbow gloves, socks, heterochromia, cleavage, heterochromia, green eyes, braid, bikini top only

As before, this lora was based on Meinaalter——

https://civitai.com/models/20945/meinaalter

if your outputs are undesirable, try this ckpt or the series of Meina XD

Notice：The characters are copyrighted by miHoYo, this lora for exchange leaning only, do not use for commercial purposes.

Please consider supporting me so I can continue to make more Models / LoRA's for the community! --> Buy Me A Coffee

This is a bit of a trash-tier LoRA since there's hardly a useable dataset for Temmie, but it produces interesting results every once in a while, so I thought I'd upload it for anyone who might be interested.

I might take a second crack at making a LoRA for Temmie in the future.

This model was trained using Furry Vixens v2.0

You can use these tags when generating images:

temmie, anthro, cat ears, temmie ears, blue and yellow striped shirt, grey hair, long hair, white fur

And be sure to use these settings!

Steps: 32

Sampler: DPM++ 2M Karras

CFG scale: 7

Size: 384x512

Model: Furry Vixens v2

Denoising strength: 0.6

Clip skip: 2

Hires upscale: 2

Hires upscaler: R-ESRGAN 4x+ Anime6B

Positive Prompt

(temmie:1.2), (chibi:1.2), (cat ears, temmie ears:1.1), paws, (blue and yellow striped shirt), grey hair, long hair, white fur, cardboard box, wooden floor, laying on back, dramatic lighting, (from above, looking at viewer), <lora:Temmie v3:0.7>

Negative prompt

(worst quality:1.4), (low quality:1.4), (easynegative:1.2)

If "Julius Zimmerman" doesn't ring any bells you're probably safe to ignore this LoRA.

Julius Zimmerman was a prolific pinup artist, creating around 17,000 pinup images before his death in 2017. His art was pencil on paper and focused primarily on fictional pop icons - Velma, Jessica Rabbit, Helen Parr, etc.

This Lora was created with kohya_ss default settings on 186 pictures with tags from r34. A LoRA with a larger image set and better curated tags would work better, but it works surprisingly well. It also works fine with character LoRAs and understands Zimmerman's more popular characters by default - Velma, Jessica Rabbit, Elvira, and probably some others.

The trigger word is "julius_zimmerman". To force a colored image use "color, colored", to force greyscale use "black_and_white, monochrome". If you do not force one or the other you'll get partially colored images. Adding realism / photorealism tags will pull the sketch towards an illustration; mixing in realistic LoRAs does the same.

There are now two versions - one Lycoris and one LoRA. If you have the Lycoris plugin I'd go with that one.

A mixture of loras I have downloaded from all over the internet. I hope it has the effect of improving the general generation of male content. You may still need a penis model (quite a few on this site) or drawings about it may not work well.

Suggested weight is ... maybe 0.8-1.1?

TO CREATE REALISTIC CONTENT I RECOMMEND

ADD DIFFERENCE MODEL + (NAI - NAISFW) first

A quick model of Lumera from FE Engage. Made with 24 images with no tag pruning. Might need some guidance to help with her cape thingy.

"CHAOS could be here" he thought, "I've never been in these wastes before. There could be CHAOS anywhere." The cool winds of magic felt good against his bare chest. "I HATE CHAOS" he thought. Where have all the gobbos gone reverberated his entire steam tank, making it pulsate even as the 9 gelt bugman's circulated through his powerful Solland veins and washed away his (merited) fear of ratmen after dark. "With a steam tank, you can go anywhere you want" he said to himself, out loud.

This lora generates chaos warriors and was trained in a booru tag style. Can do undivided or monogod (khorne, tzeentch, nurgle, slaanesh) and also chaos sorcerers. Seems to work best with AnythingV4.5 which has the side effect of making it pretty compatible with anime and memes.

This is a lazy first draft, trained on a dirty dataset of about 150 images. Including a significant amount of T-posing total war models so i recommend you put "reference pose" in negative prompts.

Below is a list of keywords used during the training of chaoswarrior:

[149] chaos warrior

[103] solo

[90] 2D

[67] simple background

[59] 3D

[58] total war

[56] undivided

[50] reference pose

[46] khorne

[42] cropped

[38] absurdly detailed

[34] group

[33] masterpiece

[26] horse

[20] chaos sorcerer

[15] archeon

[15] battle

[14] nurgle

[14] tzeentch

[13] back

[11] grayscale

[10] watermark

[9] demon

[6] slaanesh

Trained mostly on food scenes, at restaurants, but can do paperwork and reading too.

As far as lewds go? It can do flashing and implied masturbation pretty well.

For non lewds? a lot. booth seating, restaurant, library, etc. Anywhere with a desk.

Drunkoak - Artist Style 🎨

This Lora is trained alluring artworks of Letdie, make sure to also checkout his work on Twitter or Pixiv.

Usage Tips :

• regular use is at the weight of 1 to 0.9 , don't be afraid to experiment with higher values.

• adding letdie to the prompt will give the LoRA a tiny bit more influence.

• as this is a 2.5D style, it will most certainly look the best with 2.5D and semi-realistic checkpoints like "AOM A3".

Samples Settings :

• Checkpoints : AOM3 A3

• Steps : 24 to 30

• Sampler: DPM++ 2M Karras

• CFG : 7

Weight Array:

Consider posting ur creation, I love to what u make with it 😉

Please consider supporting me so I can continue to make more Models / LoRA's for the community! --> Buy Me A Coffee

This model was trained on Furry Vixens v2.0

This model is mainly trained for Anthro's, but it does small Chibi's too. It's trained for both male / female Braixens, so it should do both fairly easy.

Be sure to use the tags "gynomorph, breasts" for generating females, and "andromorph, flat chest" when you want to generate male Braixens. The tag "penis" is interchangeable for both, and there's also some training for "bukkake, many dicks" if you want to have some group fun.

It works best at a weight of 0.7, though you can use 0.4 if you want to create more generic looking furries instead.

You can use these tags when generating images:

Braixen, Braixen ears, Braixen tail, red eyes

Be sure to use these settings!

Steps: 32

Sampler: DPM++ 2M Karras

CFG scale: 7

Size: 384x512

Model: Furry Vixens v2

Denoising strength: 0.6

Clip skip: 2

Hires upscale: 2

Hires upscaler: R-ESRGAN 4x+ Anime6B

Here's an example prompt, though you can find others in the gallery.

Positive Prompt

(Braixen, Braixen ears, Braixen tail), (gynomorph:1.1), breasts, head turn, blushing, open mouth, tongue out, smile, red eyes, autumn, forest, trees <lora:Braixen v3-000006:0.7>

Negative prompt

(worst quality:1.4), (low quality:1.4), (easynegative:1.2)

I am the original creator of Lil Yunicorn style and this is an attempt to share it here. I have tried it with a few different checkpoints, TIs and LoRAs with good results so I hope you can enjoy it.

Deneme Olarak yaptığım bir lora

En iyi çıktı için ChilloutMix kullanmanızı öneriyorum (bide prompta kemal kılıçdaroğlu eklemeyi unutmayın).

Neocoill Style Lora. This is a request from my lora request form. I recommend using at 0.75 - 0.45 Strength. Faces are a little messy, to fix use Hires. fix.

Trigger word: neocoill

I created a Lora request form. Please do not submit anything that breaks the tos. Thx!

This a small character LoRa for the Overwatch Hero Kiriko.

The LoRa was trained on Galena Mix and works pretty well with it and other anime based Checkpoints. I have also been able to create some good artworks some semi-realistic Checkpoints.

20.04.2023:

The Keyword is kirkiko not kiriko.

This is so that it does not interfere with another TI I use and will be changed to smth different in the next update.

Rolo Style LoRA

This LoRA model is inspired by the art style of Rolo (@rolo_stuff) and is optimized for use with a weight between 1 and 0.5. If the output appears too overbaked, consider reducing the weight. The model was trained on Anything V3, with a focus on rolo's art style. It works particularly well with tags such as rolo, furry, body fur, tail.

Please note that while this model is not designed specifically for any particular character or subject, it is mostly trained on Rolo's main OC. As with any style model, some output may not perfectly match the desired style, but overall the model produces good results.

Please be aware that this model is not intended to be used to impersonate or steal Rolo's work. It is intended purely for fun and creative experimentation. If you use this model in your designs, please make sure to incorporate creative differences and avoid passing off the output as original work by Rolo.

To use this model in Auto1111:

1. Update the webui (use git pull or redownload it)

2. Copy the model file to stable-diffusion-webui/models/lora

3. Select the Rolo Style LoRA in the webui (make sure to include rolo, furry, body fur)

4. Adjust the weight as needed (default is 1)

Thank you for using my Rolo Style LoRA model! If you have any questions or feedback, please don't hesitate to reach out.

This is a LoRA for generating images of Persona 5 character Niijima Makoto.

Making models takes a lot of time and hard work. If you like this LoRA, please click the like button!

Recommended weights:

• Recommended weights: 0.5~0.7

  (Note: If the weight is below 0.4, the LoRA may not work effectively.)

這是一個用於生成女神異聞錄5角色新島真的圖像的LoRA

製作模型需要花費很多的時間和精力。如果您喜歡這個LoRA，請幫忙按一個“讚”！

建議權重：

• 建議權重：0.5〜0.7

  （注意：如果權重低於0.4，則LoRA可能無法生效。）

2023,04,21 v2.0

三个衣服能正常画出来了

泳衣：Dorothy-sw, swimsuit

私服：Dorothy-pc, dress

女仆：Dorothy-en, enmaided

Google Translate:

The three clothes can be drawn normally

swimsuit：Dorothy-sw, swimsuit

Private Clothes：Dorothy-pc, dress

enmaided：Dorothy-en, enmaided

OLD v1.0

只练了水手服

当你将6个女主的lora强度0.15发动融合，你就会获得八宝备仁画风lora

I was planning to do a Cody Travers lora for a long time, then I realized that today (Apr 18) is his birthday. I still need to do some tests and I want to post a version with Mayor and Prisoner. But for now, enjoy the classic Cody from first Final Fight (my fave version 30 years later)

keyword is CodyTravers

works with Classic, white t-shirt, ripped shirt, shirtless... Although my database had a lot of his final fight look, t-shirt wants to disappear all the time. I need to test more, I just wanted to post on his birthday :P

UPDATE

0.9 is a version where the Mayor and Prison clothes were included. Mayor usually will be triggered by the words mayor, vest, tie, shirt, pants, slick hair. The prison jumpsuit is a mystery. This is why I am not considering the 1.0 yet lol.

The Juice is in the description below

Tips to use:

• Use LoRA weight between 0.6-0.8

• To summon her use the trigger Tags plus following tags: solo, long hair, brown hair, blue eyes, necktie, red necktie, jacket, pantyhose, shirt, looking at viewer, shorts, white shirt, blush, red hair, legwear under shorts, black shorts, smile, belt, short shorts, off shoulder, black pantyhose,

We used one of Meina's model to generate the images.

If you have any questions feel free to ask!

Hope you guys like it! ❤️

You can ask for character LoRAs on our patreon page. Consider supporting us. 😘

A LoRA of Koizumi Mahiru from Danganronpa 2. Trained on NAI.

Works well with a strenght of 0.8-1. See samples for good prompts.

This Lora affect those prompts : necktie, plaid necktie, shirt, looking at viewer, short sleeves, green dress, white shirt, pinafore dress, collared shirt, skirt, pleated skirt, black socks, black footwears, dress.

If you like this LoRA, please let a review! I'm planning to do every female Danganronpa characters so tell me who you wants first!

All of my LoRAs will always be free. If you paid for this or an image using this LoRA, you got scamed! If you wants to supports me, feel free to buy me a Ko-Fi.

more model in my huggingface - https://huggingface.co/myka/lora-jibril-no-game-no-life

Description of rare models:

FULL - model trained on all available images on the Internet. (390 img)

Note #1: Wanted to try to mix it with my in development, quite NSFW, Mind Flayer Sex Locon but TOS wouldn't allow it to be uploaded. Check it out on my twitter. If the Locon won't be able to be uploaded here once it's ready, I'll have to find somewhere else to put it. Recommendations appreciated!

What does it do you ask? The songstress from across time, Vivy, is ready to sing on your stable diffusion build and potentially... Well. I shouldn't give spoilers to the show. It's great, watch it! And it actually finishes in a season! Also, Vivy is just a dope character in general and one helluva fighter.

How to use it:

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

First we have the core trigger baked into the Locon.

VivyA - Main trigger word. And only one. Can remove if you want to help make other outfits.

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

Picture Descriptions:

1-2: AnythingV3 using just this Lora.

3: AnythingV3 combo'd with my in progress "Sit on/Ass Focus" Locon.

4: AnythingV3 combo'd with my in progress Racer Queen Locon.

5: AnythingV3 combo'd with my in progress Jiangshi Transformation Locon.

6: AnythingV3 combo'd with my in progress style of Defaultz17 Locon.

7: AnythingV3 combo'd with my released, super beta, Phone Pic Locon.

8: AbyssOrange using just this Lora.

9: Perfect orld using just this Lora.

10: RevAnimated using just this Lora.

How was it trained?

It was trained on animefull_final with 22 images with the usual self-pruning and tagging and a wide range of different images. I like keeping my prompts simple for the first few pictures so that you can see it without a whole bunch of mumbo jumbo. Make your pictures better than mine though!

What's next for this?

Should be good to go with this one but let me know of errors.

What's next that isn't this?

More character Loras. (Viper and Rupee from Nikke, Spectre Tenebria from E7, Zephia from Fire Emblem Engage, Ivy from Soul Calibur 5, Heine from Combatants Will Be Dispatched, Conqueror Lilias from E7, Evelynn from LoL (both base and All Out version), Tharja from Fire Emblem, Laharl-Chan from Disgaea 5, Miku from Idaten Deities, Kuribayashi Shino from GATE, Rin Sohma from Fruits Basket, Kasumi from P5, Ann from P5, Ikumi Mito from Food Wars, Alice Nakiri from Food Wars, Chika Fujiwara from Kaguya-sama, Lulu from FF10, Momo from My Hero, Sheik from Zelda, Lucoa from Dragon Maid, Laharl-chan from Disgaea, Dominique de Sade from Vanitas, and Dos from One Punch Man.) And maybe Wilhelm van Astrea, my favorite Re:Zero husbando or a Prinny from Disgaea.

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

I also have some more quite NSFW Locon's. I decided to list them as well, so if people have preferences on which to fix up first let me know. Specifically, Bound Yuri (two girls tied up, breast to breast and likely kissing), Sultry Succubus Transformation, Tied up with vibrators (basically the waifu tied up with vibrators and being overwhelmed with pleasure), Beta Censorship (Automatically censors any images you create to make them BetaSafe), Demonic Transformation, being stepped on, Robotification, Defeated (decently NSFW defeat scenario for the waifu getting abused), Petrified (I'm unsure if I should bother with this one anymore), Infection (this one is VERY nsfw, as the waifu is completely infected), Dominatrix (waifu becoming a dominatrix), Brainwashing Machine, Tentacles, Pillory, Masked Evil (A mask that possess' the waifu), Latex Corruption, Combat Drone (hard to describe but basically making them a mindless sci-fi combat warrior), JAV Supervillain, Jiangshi (Transforms into a Jiangshi), Mind Flayer Defeat, Blacked, Hypnotized Harem, Hourglass Figure (just to make your waifu curvy), and Holding a Leash.

And, uh, I want to make a few SFW Locon's so I have a Goth Chick, Delinquent (The delinquent girls in japanese schools specifically), GoodCop/BadCop (Locon with outfits for being a good and for being a bad cop), Maid (both normal and naughty variants), Biker Chick (ride a motorcycle and look like you should), Race Queen (Chiling sexily on a car), forest fairy, Anime Officer, barbarian.

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

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

Hello friends, this is my first LoRA and I'm glad to introduce Lulu from Final Fantasy X to you.

This is NSFW model with mostly open dressed, lingerie and naked variants.

LoRA was trained on Perfect world model with 3D NSFW/SFW arts.

54 arts were used with 75 repeats and 1 epoch.

Also I have version with 8 epoches but I think version with one epoch is better.

0.6 strength of Lora is the best choice.

Some necessary booru prompt for better experience:

1girl, black_hair, beads, hair_ornament, hair_over_one_eye, hair_stick

If you have troubles with distorted eyes:

• use highres.fix (I prefer not to do this);

• use SD Upscale script in img2img (For me 1.25-1.5x with AnimeSharp or UltraSharp is good on 0.2-0.3 denoising strengh);

• use extention "dddetailer" (You can find it on github and instrall directly in extension tab by copying .git url) It works well like inpainting. Use 0.3 denoising str, copy your promts and neg promts and pick dd-person or anime face model in detection model A);

• Use inpaint (but with SD Upscale and dddetailer I don't need this in 90% times);

• Use Photoshop. It is necesarry all time if you want to make your art better. I didn't use it on preview to save prompts and help somebody who need prompts.

• Also use prompt "brown eyes". It is more balanced than "red eyes".

I will be glad to receive feedback from you and a demonstration of your art with my LoRA!

Trained some fanarts and ingame screenshots on MeinaHentai model

use hirefix for better results

0.8 and bellow is good

example prompt:Antimage_Wei, 1girl, single_thighhigh, cherry_blossoms, falling_petals, long_hair, petals, solo

for the weapon try adding: dual_wielding, holding, holding_weapon

To avoid the problem of improper commercial or illegal use disputes, when you start using the model, you are fully aware of the declaration document. I do not support the use of this model to generate adult content. The above model is for learning purposes only. Users must abide by local laws and regulations when using the model, and must not violate others' honor, privacy and portrait rights. I solemnly declare that the final right of interpretation of this model belongs to expired Rice Noodle Meow myself. If you have commercial needs, please directly contact with business through the official Weibo account of the owner of portrait rights. Without authorization, no commercial use of this model is allowed, otherwise all consequences will be taken by yourself. This declaration document does not represent the expired Rice Noodle Meow official, nor does it represent the expired Rice Noodle Meow itself, but is only used for disclaimer by software author himself.

为避免不正当的商用或非法使用纠纷问题，当您开始使用该模型即表明您已对此声明文件充分了解。本人并不支持使用这种模式来生成成人内容。以上模型仅供学习之用。用户在使用该模型时，请务必须遵守当地的法律法规，不得侵犯他人的荣誉、隐私和肖像权。本人严正声明该模型的最终解释权归 过期米线喵本人所有，若您有商用需求请直接通过肖像权归属者官方微博等渠道取得商用联系，未经授权者一律不允许将其商用，否则一切后果自负。该声明文件不代表过期米线喵官方，亦不代表过期米线喵本人，仅用于模型作者本人免责申明。

부적당한 상업성이나 불법 사용으로 인한 분쟁을 피하기 위해, 모델을 시작할 때 제보 파일을 숙지해 두십시오.저는이 모델을 성인용 컨텐츠를 만드는 데 사용하지 않습니다.위의 모델은 학습에만 사용될 수 있다.사용자는 모델을 사용할 때 현지 법률 및 규정을 준수해야 하며 타인의 명예, 사생활, 초상권을 침해해서는 안 된다.본 모델의 최종해석권은 초상권 본인에게 있음을 정중히 선언합니다.만약 상업적인 요구가 있다면 초상권자의 웨이보 공식 계정을 통해서 단위와 직접 연락하세요.허락 없이, 모델은 상업적으로 사용될 수 없으며, 그렇지 않으면 모든 결과는 모델 사용자가 책임져야 한다.이 선언문은 초상권자인 관변단체나 초상권자 자체를 대표하지 않으며 소프트웨어 작성자 본인의 고지사항에만 사용된다.

このモデルを使い始めるということは、不正な商用利用や不正利用の問題を回避するために、この宣言文を十分に理解しているということです。このモデルを使ったアダルトコンテンツの生成はサポートしていません。ただの学習用モデルです。ユーザーがこのモデルを利用する際には、現地の法律に従わなければなりません。他人の名誉、プライバシー、肖像権を侵害してはなりません。私は厳正に声明の本モデルのファイルの最終的な解釈の権利が期限切れの米線猫本人が所有して、もしあなたは商用の需要があって直接肖像権の帰属の方の公式の微博などのチャンネルを通じて商用の連絡を取得して、許可を得ていません許可しないで商業利用目的です，さもなければすべての結果の利用者です責任を負います。この声明の文書は、米線の公式の代表していませんし、米線の代表していません本人は、モデルの作者本人の免責を声明すのされます。

Resource used: https://civitai.com/models/22530

This was requested through Google Forms

Weight: 0.8

Trigger + Appearance: som gohan, black hair, spiked hair, black eyes

Sample Outfit: dougi

Super Saiyan: super saiyan, blonde hair, spiked hair, blue eyes

Trained of mj images for the prompt

van goghs a starry night over new york city, recursive network graphs within neural networks made of stars and galaxies in space, moebius + pascal campion, incredibly detailed + sharpen + professional lighting

Resource used: https://civitai.com/models/22530

This was requested through Google Forms

Weight: 0.7

Trigger + Appearance: ritz malheur, red hair, green eyes

Outfit: armor, thighhighs

She can properly hold a staff if you're lucky.

Resource used: https://civitai.com/models/22530

This was requested through Google Forms

Weight: 0.7

Trigger + Appearance: aliceliese, hair ribbon

Outfit: dress

Future princess

guardian tales character

you can active with:future princess

for better result you can add :blonde hair, armor, hair band, breastplate,shoulder armor, light smile,bangs , ,boots ,, green scarf ,armored boots,, hair over one eye ,blue eyes ,gloves , ,pantyhose

if you want holding shield and wepon then add :holding,shield, polearm, scythe, staff, weapon, (if dont want it then add it to negative prompt)

you can find me on :https://www.pixiv.net/en/users/34515535

or can support me on :https://www.buymeacoffee.com/bessu216z

Liang Xing Style Lora. This is a request from my lora request form. I recommend using at 0.75 - 0.45 Strength. Faces are a little messy, to fix use Hires. fix.

Trigger word: liang xing

I created a Lora request form. Please do not submit anything that breaks the tos. Thx!

Resource used: https://civitai.com/models/22530

This was requested through Google Forms

Weight: 0.7 - 0.9

Trigger + Appearance: plachta, light purple hair, hair between eyes, hair ornament

Outfit: dress, white thighhighs

0.8 best

learn from 4044 majic his telegram group is: https://t.me/majicplayground

my channel is https://t.me/dollsgirlswifes

Based in US, Uchi, Data Scientist

dm here https://t.me/happygoat1946

Workstation: NVIDIA® A40 (from lab)

#ai #stablediffusion #chilloutmix #lora

Resource used: https://civitai.com/models/22530

This was requested through Google Forms

Weight: 0.7 - 0.9

Trigger + Appearance: cornelia \(atelier\), hair ornament

Outfit: hat, off shoulder, japanese clothes, sleeves past wrists, sleeves past fingers

v1.0

Other versions

A LoRA based on 4267 Grimoire A: Private Grimoire Magic Academy cards, recognizing 60+ characters to varying degrees.

A request by Civit user themonky007. The model is far from complete but there are currently no plans to further improve it. The dataset is available for those willing to take on the task.

Lesson of the day: suitable learning rates can differ significantly between datasets (this uses 2e-4 while the Idolmaster models use 1e-3).

An undertrained version (v0.1) with card text is also available. Unlike v1.0, v0.1 uses Japanese names with spaces (e.g. 朝比奈 龍季) as character tags. Interestingly, it can produce the correct Romaji initials for out-of-distribution characters quite reliably.

Usage

The style itself doesn't have a trigger word. For the characters, below are the character tags and respective top 20 related tags.

Minami Tomoka,
brown hair, short hair, blush, brown eyes, 1girl, ahoge, smile, looking at viewer, bow, solo, open mouth, hair bow, skirt, :d, outdoors, ribbon, dress, necktie, sky, thighhighs

Nobara Hime,
blonde hair, flower, long hair, 1girl, hair ornament, blue eyes, hair flower, blush, solo, rose, drill hair, looking at viewer, smile, open mouth, frills, bow, red flower, red rose, dress, skirt

Kannagi Rei,
1girl, long hair, brown hair, solo, brown eyes, looking at viewer, blush, hair between eyes, smile, skirt, outdoors, japanese clothes, holding, open mouth, black hair, yellow eyes, bangs, long sleeves, weapon, day

Shinonome Aira,
long hair, 1girl, looking at viewer, solo, red eyes, blush, smile, open mouth, fang, dress, multicolored hair, grey hair, frills, purple hair, thighhighs, hat, :d, hair between eyes, outdoors, long sleeves

Fuyuki Noel,
1girl, blonde hair, green eyes, solo, braid, blush, looking at viewer, bow, open mouth, smile, short hair, hair bow, outdoors, twin braids, skirt, :d, ribbon, sky, day, long hair

Kishida Natsumi,
1girl, long hair, brown hair, twintails, solo, brown eyes, blush, ahoge, open mouth, looking at viewer, smile, hair ornament, ribbon, camera, hair between eyes, holding, fang, skirt, red hair, :d

Fuyuki Eve,
blonde hair, green eyes, long hair, 1girl, solo, hairband, blush, looking at viewer, skirt, open mouth, smile, dress, gloves, outdoors, long sleeves, holding, shirt, flower, sitting, thighhighs

Minazuki Fuko,
blonde hair, blue eyes, long hair, 1girl, braid, solo, twintails, ribbon, hair ribbon, blush, looking at viewer, thighhighs, smile, skirt, open mouth, gloves, necktie, very long hair, school uniform, hat

Rurikawa Akiho,
brown hair, blush, 1girl, brown eyes, solo, open mouth, looking at viewer, ribbon, smile, skirt, hair ribbon, bow, :d, short hair, dress, twintails, outdoors, socks, long hair, day

Tachibana Ui,
long hair, 1girl, purple eyes, braid, solo, looking at viewer, antenna hair, ribbon, blonde hair, skirt, hair between eyes, bow, hair ribbon, blush, hair bow, sitting, pantyhose, dress, outdoors, flower

Momose momo,
pink hair, 1girl, solo, blush, ahoge, hair ornament, looking at viewer, smile, side ponytail, open mouth, short hair, skirt, brown eyes, long hair, :d, flower, shirt, holding, pink eyes, outdoors

Moriya Tsukuyo,
green eyes, 1girl, long hair, solo, blush, ribbon, looking at viewer, aqua hair, hair ribbon, open mouth, green hair, two side up, skirt, thighhighs, smile, hair ornament, long sleeves, holding, outdoors, twintails

Tateno Nozomi,
1girl, long hair, solo, black hair, purple eyes, blush, open mouth, looking at viewer, smile, skirt, purple hair, thighhighs, sitting, black eyes, gloves, very long hair, white gloves, boots, outdoors, striped

Hattori Azusa,
green hair, 1girl, green eyes, long hair, solo, looking at viewer, blush, smile, open mouth, scarf, hair ornament, gloves, pantyhose, outdoors, japanese clothes, ninja, fingerless gloves, shorts, red scarf, sparkle

Minase Kaoruko,
long hair, purple eyes, 1girl, black hair, breasts, solo, looking at viewer, smile, mole under eye, mole, ponytail, hair between eyes, blush, open mouth, very long hair, large breasts, skirt, outdoors, day, long sleeves

Otonashi ritsu,
short hair, 1girl, solo, blush, brown hair, smile, yellow eyes, open mouth, brown eyes, looking at viewer, fingerless gloves, gloves, red hair, shorts, thighhighs, skirt, necktie, shirt, bangs, :d

Yusa Naruko,
short hair, 1girl, purple eyes, solo, black hair, smile, looking at viewer, blush, shirt, purple hair, pants, jewelry, open mouth, holding, outdoors, skirt, gloves, white shirt, long sleeves, sky

Shirafuji Konoha,
long hair, blue hair, 1girl, green eyes, hair ornament, blush, heart, solo, looking at viewer, smile, heart hair ornament, open mouth, mole under eye, mole, hairclip, skirt, thighhighs, outdoors, shirt, very long hair

Takanashi Miyu,
short hair, 1girl, yellow eyes, solo, looking at viewer, smile, grey hair, blush, open mouth, flower, outdoors, hair between eyes, hairband, breasts, :d, shirt, day, brown eyes, gloves, long sleeves

Narumi Jun,
blue hair, short hair, 1girl, solo, blush, looking at viewer, smile, open mouth, yellow eyes, earrings, jewelry, green eyes, breasts, skirt, outdoors, necktie, shorts, sitting, shirt, school uniform

Shiina Yukari,
purple hair, glasses, 1girl, long hair, solo, braid, smile, brown eyes, looking at viewer, open mouth, breasts, blush, bow, twin braids, skirt, very long hair, outdoors, dress, holding, flower

Wagatsuma Asari,
blue hair, blush, blue eyes, short hair, solo, looking at viewer, 1girl, open mouth, smile, thighhighs, ahoge, bow, 1boy, hat, skirt, male focus, white thighhighs, otoko no ko, frills, crossdressing

Nakazuki Sara,
brown hair, 1girl, brown eyes, ahoge, blush, solo, open mouth, long hair, bow, smile, animal ears, dress, thighhighs, looking at viewer, frills, outdoors, :d, rabbit ears, gloves, hair ornament

Sumeragi Ayaka,
pink hair, 1girl, blush, looking at viewer, smile, blue eyes, solo, open mouth, short hair, hair ornament, skirt, jewelry, long hair, earrings, bangs, purple eyes, star, :d, holding, gloves

Tsukimiya Sana,
blue eyes, 1girl, grey hair, solo, blush, looking at viewer, short hair, long hair, smile, dress, holding, breasts, long sleeves, open mouth, apron, bangs, indoors, shirt, gloves, outdoors

Kusunogi Arisu,
long hair, blue eyes, blonde hair, multiple girls, 2girls, blush, bow, open mouth, hair bow, dress, red eyes, smile, twintails, grey hair, ribbon, looking at viewer, very long hair, blue hair, hat, skirt

Kurusu Homura,
black hair, 1girl, solo, short hair, looking at viewer, blush, red eyes, shorts, open mouth, bangs, hood, outdoors, shirt, skirt, gloves, navel, jewelry, fingerless gloves, belt, short shorts

Mamiya Chika,
brown hair, 1girl, brown eyes, blush, long hair, solo, bow, hair bow, looking at viewer, open mouth, smile, jewelry, ponytail, breasts, skirt, medium breasts, sitting, navel, outdoors, side ponytail

Reisen Aoi,
1girl, brown hair, solo, blush, long hair, hair ornament, flower, single hair bun, smile, hair bun, looking at viewer, open mouth, hair flower, bangs, blue eyes, japanese clothes, aqua eyes, skirt, holding, sidelocks

Emilia Bloomfield,
long hair, 1girl, solo, green eyes, blonde hair, blush, looking at viewer, smile, open mouth, ribbon, skirt, outdoors, hair ribbon, gloves, sky, holding, day, :d, bow, shirt

Kirizuka Moegi,
black hair, long hair, 1girl, blush, braid, solo, looking at viewer, twintails, ribbon, hair ribbon, twin braids, open mouth, pink eyes, very long hair, red eyes, skirt, long sleeves, bangs, hair between eyes, smile

Koshimizu Sophia,
blonde hair, 1girl, blue eyes, long hair, solo, blush, open mouth, smile, bow, hair rings, looking at viewer, hair bow, twintails, :d, skirt, outdoors, ribbon, day, gloves, shirt

Mary Williams,
blonde hair, blue eyes, 1girl, solo, long hair, breasts, smile, looking at viewer, navel, midriff, jacket, weapon, open mouth, outdoors, shorts, gun, belt, medium breasts, blush, shirt

Hikawa Saki,
blue hair, long hair, 1girl, brown eyes, blush, solo, hair ornament, yellow eyes, open mouth, hairclip, side ponytail, looking at viewer, skirt, shirt, blue skirt, long sleeves, breasts, outdoors, necktie, white shirt

Nanjo Ren,
purple eyes, 1girl, blush, solo, black hair, long hair, looking at viewer, smile, open mouth, bangs, purple hair, skirt, japanese clothes, holding, :d, ribbon, hair ornament, shirt, outdoors, day

Ebina Ayase,
long hair, 1girl, green eyes, solo, pink hair, blush, smile, looking at viewer, braid, open mouth, breasts, outdoors, skirt, shirt, dress, :d, day, gloves, sitting, flower

Yuki Sena,
green hair, glasses, short hair, 1girl, solo, green eyes, looking at viewer, open mouth, yellow eyes, semi-rimless eyewear, skirt, smile, blush, outdoors, hat, breasts, pantyhose, hair ornament, black-framed eyewear, necktie

Charlotte Dior,
blonde hair, long hair, 1girl, blue eyes, solo, blush, looking at viewer, breasts, open mouth, large breasts, smile, holding, dress, very long hair, long sleeves, nun, cross, habit, skirt, shirt

Saibara Yueko,
purple hair, short hair, 1girl, solo, purple eyes, blush, braid, open mouth, hair ornament, smile, looking at viewer, skirt, pantyhose, outdoors, black pantyhose, necktie, :3, jewelry, star, sky

Kazeyari Mina,
1girl, solo, eyepatch, long hair, grey hair, looking at viewer, open mouth, blue eyes, hair ornament, blush, smile, cape, skirt, two side up, thighhighs, bow, aqua eyes, long sleeves, hair between eyes, black thighhighs

Takeda Torachiyo,
blonde hair, long hair, 1girl, solo, looking at viewer, smile, brown eyes, breasts, yellow eyes, gloves, medium breasts, midriff, skirt, fingerless gloves, open mouth, navel, blush, black gloves, jacket, hair between eyes

Nabatame Tsukasa,
short hair, 1girl, solo, purple eyes, blue hair, smile, looking at viewer, bandages, navel, shorts, midriff, breasts, green hair, ahoge, gloves, skirt, open mouth, pink eyes, outdoors, sitting

Rurikawa Haruno,
brown hair, long hair, 1girl, solo, blush, ribbon, red eyes, open mouth, smile, hair ribbon, earrings, looking at viewer, jewelry, breasts, brown eyes, skirt, long sleeves, jacket, shirt, heart

Tokisaka Chitose,
black hair, blue eyes, 1girl, long hair, solo, looking at viewer, smile, hair between eyes, blush, open mouth, skirt, breasts, sky, very long hair, long sleeves, sitting, bangs, medium breasts, dress, thighhighs

Futami Kokoro,
pink hair, 1girl, short hair, solo, purple eyes, blush, open mouth, looking at viewer, skirt, thighhighs, smile, sky, gloves, blue skirt, outdoors, school uniform, black thighhighs, fingerless gloves, closed eyes, breasts

Yonamise Rina,
purple hair, 1girl, blue eyes, solo, short hair, blush, open mouth, looking at viewer, smile, :d, breasts, bandaid, skirt, shirt, bandaid on face, outdoors, water, sky, midriff, navel

Ellen Amedick,
red hair, 1girl, yellow eyes, solo, eyepatch, looking at viewer, breasts, smile, long hair, blush, gloves, short hair, braid, outdoors, pants, weapon, holding, white gloves, flower, brown eyes

SHIshido yuki,
purple eyes, short hair, blue hair, 1girl, solo, glasses, looking at viewer, thighhighs, open mouth, skirt, sitting, headphones, jacket, smile, hood, dress, school uniform, multiple girls, breasts, holding

Asahina Tatsuki,
blue eyes, long hair, brown hair, 1girl, solo, blush, ahoge, looking at viewer, shirt, smile, open mouth, jewelry, hair between eyes, breasts, pants, outdoors, bangs, skirt, bracelet, belt

Yukishiro Mashiro,
1girl, long hair, blue hair, solo, blue eyes, looking at viewer, smile, breasts, blush, very long hair, dress, open mouth, jewelry, skirt, cleavage, medium breasts, hair between eyes, earrings, outdoors, braid

Soma Rena,
1girl, solo, green eyes, open mouth, fang, blush, short hair, smile, grey hair, looking at viewer, dark skin, animal ears, outdoors, dark-skinned female, :d, breasts, long hair, sky, day, barefoot

Lee Xiaolian,
green eyes, green hair, 1girl, blush, hair bun, solo, smile, open mouth, looking at viewer, bun cover, dress, double bun, short hair, single hair bun, chinese clothes, holding, long hair, shorts, outdoors, food

Qiao Mingling,
long hair, braid, 1girl, blue eyes, solo, blush, smile, open mouth, pink hair, looking at viewer, red hair, single braid, ahoge, chinese clothes, fang, dress, :d, china dress, ribbon, gloves

Kisaragi Sora,
1girl, solo, long hair, green eyes, red hair, goggles, blush, goggles on head, gloves, open mouth, looking at viewer, skirt, thighhighs, smile, ahoge, breasts, jewelry, shirt, black thighhighs, earrings

Hasekura Toko,
1girl, long hair, solo, black hair, brown eyes, ponytail, looking at viewer, blush, very long hair, brown hair, open mouth, breasts, japanese clothes, skirt, smile, bangs, weapon, flower, outdoors, hair ornament

Jinguji Hatsune,
long hair, blue eyes, 1girl, solo, blush, brown hair, looking at viewer, smile, open mouth, hairband, ponytail, skirt, red hair, frills, ribbon, very long hair, thighhighs, :d, long sleeves, bow

Yayoi Ruqa,
braid, 1girl, solo, orange hair, smile, green eyes, open mouth, blush, looking at viewer, bird, breasts, yellow eyes, medium breasts, long hair, outdoors, hair ornament, gloves, cleavage, skirt, boots

Satonaka Karin,
brown eyes, brown hair, 1girl, solo, blush, short hair, open mouth, looking at viewer, smile, braid, flower, breasts, freckles, sky, outdoors, twin braids, :d, hair ornament, bow, skirt

Madokano Marika,
black hair, green hair, multicolored hair, short hair, 1girl, looking at viewer, green eyes, two-tone hair, solo, open mouth, smile, breasts, blush, outdoors, blue eyes, jacket, mask, sky, shorts, hat

Urajiro Nana,
black hair, 1girl, long hair, blush, solo, red eyes, ahoge, open mouth, looking at viewer, smile, skirt, braid, bow, bangs, hair bow, :d, brown eyes, outdoors, holding, ribbon

Matsushima Michiru,
short hair, 1girl, brown hair, green eyes, solo, open mouth, blush, looking at viewer, smile, hairband, skirt, :d, shirt, jacket, hair ornament, holding, indoors, gloves, outdoors, breasts

Inukawa Nene,
1girl, pink hair, long hair, solo, blush, hair ornament, bell, red eyes, smile, looking at viewer, open mouth, ribbon, pink eyes, fang, very long hair, jingle bell, dress, hair bell, hair between eyes, :d

Nanayoshi Chihiro,
short hair, solo, 1girl, blush, bow, brown eyes, green hair, open mouth, smile, hair bow, looking at viewer, hair ornament, heart, blonde hair, skirt, red bow, :d, school uniform, sky, outdoors

Lizette,
black hair, solo, purple eyes, 1girl, multicolored hair, hair between eyes, two-tone hair, looking at viewer, white hair, long hair, 1boy, blush, bangs, school uniform, skirt, necktie, short hair, male focus, indoors, breasts

Uno,
1girl, smile, pink hair, twintails, long hair, solo, grey eyes, looking at viewer, hair ornament, blush, open mouth, holding, dress, thighhighs, :d, confetti, multicolored hair, striped, stuffed toy, one eye closed

Training info

Dataset:

• 4267 Grimoire cards extracted from official offline apps with text removed (characters labeled by themonky007)

Tags generated with SmilingWolf/wd-v1-4-convnextv2-tagger-v2 with a threshold of 0.35. No balancing and tag cleaning. Images were processed with RealESRGAN_x4plus_anime_6B.

Inspired by a 2022 paper, a custom LR scheduler was used:

SequentialLR(optimizer, [
    LinearLR(optimizer, 0.1, total_iters=steps_per_epoch),
    CosineAnnealingWarmRestarts(optimizer, steps_per_epoch*2, T_mult=2)
], [steps_per_epoch])

Training cost: ~18 T4-hours

Data distribution:

             character  cards
1        Minami Tomoka     94
2          Nobara Hime     86
3          Kannagi Rei     82
4       Shinonome Aira     80
5          Fuyuki Noel     80
6      Kishida Natsumi     79
7           Fuyuki Eve     78
8        Minazuki Fuko     78
9       Rurikawa Akiho     77
10        Tachibana Ui     76
11         Momose momo     75
12      Moriya Tsukuyo     74
13       Tateno Nozomi     74
14       Hattori Azusa     74
15      Minase Kaoruko     72
16      Otonashi ritsu     72
17         Yusa Naruko     72
18    Shirafuji Konoha     72
19      Takanashi Miyu     72
20          Narumi Jun     70
21       Shiina Yukari     70
22     Wagatsuma Asari     70
23       Nakazuki Sara     70
24      Sumeragi Ayaka     68
25      Tsukimiya Sana     68
26      Kusunogi Arisu     68
27       Kurusu Homura     68
28        Mamiya Chika     68
29          Reisen Aoi     68
30   Emilia Bloomfield     68
31      Kirizuka Moegi     68
32    Koshimizu Sophia     68
33       Mary Williams     68
34         Hikawa Saki     66
35           Nanjo Ren     66
36         Ebina Ayase     66
37           Yuki Sena     66
38      Charlotte Dior     66
39       Saibara Yueko     66
40       Kazeyari Mina     66
41    Takeda Torachiyo     66
42    Nabatame Tsukasa     66
43     Rurikawa Haruno     64
44    Tokisaka Chitose     64
45       Futami Kokoro     64
46       Yonamise Rina     64
47       Ellen Amedick     62
48       SHIshido yuki     62
49     Asahina Tatsuki     62
50   Yukishiro Mashiro     62
51           Soma Rena     62
52        Lee Xiaolian     62
53       Qiao Mingling     62
54       Kisaragi Sora     60
55       Hasekura Toko     60
56     Jinguji Hatsune     60
57          Yayoi Ruqa     60
58      Satonaka Karin     58
59     Madokano Marika     54
60        Urajiro Nana     50
61  Matsushima Michiru     50
62        Inukawa Nene     48
63   Nanayoshi Chihiro     44
64             Lizette      8
65                 Uno      4