这是用于生成胶片风格AI写真照片的LoHA & LoRA模型。与majicMIX realistic、MoonMIX或Chilloumix搭配使用，可以生成逼真的胶片风格照片。同时推荐搭配TEXTUAL INVERSION模型：bad_pictures、negative_hand Negative Embedding。这两个Negative Embedding模型不会显著影响到胶片的质感。

This is LoHA & LoRA models for generating film style AI portrait photos. when used with majicMIX realistic、 MoonMIX or Chilloumix, it can generate realistic film style photos. At the same time, it is recommended to use the TEXTUAL INVERSION model: bad_pictures, negative_hand Negative Embedding. These two Negative Embedding models will not significantly affect the film texture.

由于新版本的Provia模型太过真实，我上传的好多效果图被系统判为疑似真实照片给夹了。。我只能先AI生成些离谱内容的预览图，不然过不了审。。。

Because the effects of the new Provia model are too realistic, many of the effect pictures I uploaded were judged by the system as suspected real photos and were rejected. I can only first generate some absurd preview pictures...

2023.5.19 Introducing "the Provia Version: FilmGX2 LoHA model "

Provia版本不同于之前0.1到4.0版本LoRA模型，Provia是全新的胶片模拟LoHA模型，使用60张优选后的胶片照作为训练集，以及505张胶片照作为正则化图像集训练而来。

The Provia version is different from the previous Lora models from 0.1 to 4.0. Provia is a brand new film simulation Loha model, trained with 60 selected film photos as the training set and 505 film photos as the regularization image set.

使用Loha模型的一种方法:

请安装以下插件，https://github.com/KohakuBlueleaf/a1111-sd-webui-locon，然后调用方法与Lora类似，需要将模型放入...\models\Lora文件夹中，随后使用<lora:FilmProvia2:XXX>指令调用。

另一种使用Loha模型的方法: → https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris

How to use the Loha model:

Please install the following plugins, https://github.com/KohakuBlueleaf/a1111-sd-webui-locon, then use the same method as Lora. You need to place the model in the \models\Lora folder, and then call it with the <lora:FilmProvia2:XXX> command.

Another way to use the Loha model → https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris

Provia版本的胶片质感要明显强于1.0到4.0版本，面部正确率要明显强于0.1和0.2版本。如果你非常在意胶片真实质感，那这个版本是整个胶片系列模型中最推荐使用的版本。但请注意，因为这个版本没有融合其他模型进行面部优化，因此面部崩坏的概率会比v4高一些。

The film texture of the Provia version is significantly stronger than versions 1.0 to 4.0, and the face accuracy rate is significantly stronger than versions 0.1 and 0.2. If you are very concerned about the real film texture, this version is the most recommended in the entire film series model. But please note that because this version does not integrate other models for facial optimization, the probability of facial collapse will be higher than v4.

请注意：

1. Provia版本不在需要Trigger Words。

2. 我常用的CFG范围是6~7，Lora权重通常在0.6~0.8。

3. Clip skip设为1或2时，生成的效果会有所不同，大家可以多尝试看自己喜欢哪种风格。

4. 本LoHA模型的训练集主要为亚洲胶片人像照片，生成欧美人像的效果会受影响。

Please note:

1. The Provia version no longer requires Trigger Words.

2. My commonly used CFG range is 6 to 7, and the Lora weight is usually 0.6 to 0.8.

3. When the Clip skip is set to 1 or 2, the generated effects will be different. You can try more to see which style you like.

4. The training set of this LoHA model is mainly Asian film portrait photos, and the effect of generating European and American portraits will be affected.

hayakawa taduna (umamusume) Lora (LYCORIS - loha) ver.2

※Please do not post NSFW content in places where it can be seen by many people, such as on social media.

I promise! 😁👍

※NSFWはSNSなど、多くの人目につく場所に公開しないようお願いします。

約束だよ😁👍

I'd be happy if you could follow and support me to increase my motivation.

モチベーション向上のためにフォローや支援してくれると嬉しいぞ。

👇😎👇

https://twitter.com/oLAcle1220

https://patreon.com/olacle

Trigger words

• base

  • Trigger word

    • hayakawa_taduna

  • custom prompt (chara)

    • brown hair,long hair,split ponytail,green eyes

    • horse girl,horse tail,horse ears

  • custom prompt (secretary clothes)

    • bowler hat,green headwear,green jacket,white shirt,double-breasted,collared shirt,pencil skirt,green skirt,green footwear,yellow necktie,yellow ribbon,yellow bow,buttons,horseshoe ornament,black pantyhose,wristwatch,

Recommended settings（おすすめ設定）

weight 0.45~

If you want to show the ears, please use the prompt "horse girl, horse tail, horse ears".

To dress the character in other clothing, please adjust the weight accordingly.

耳を出したい場合などは「horse girl,horse tail,horse ears」のプロンプトを使用してください。

その他の服を着させたい場合はweight を調整してください。

LyCORIS trained on Rocket from
Guardians of the Galaxy

Images given wern't cherry picked, so you can get MUCH better results

Can't wait to see your generations^^

LoHA trained on 3 skins of Fiora from League of Legends

Default

Headmistress

Pool Party

If you have any suggestions for who I should do next, please let me know.

models used

anime: CoffeeMix, PrimeMix, Anything and Everything

2,5D: MeinaUnreal - V3, PsyGon Animerge - v1.1, HELLmix, ReV Animated

realistic: fantasticmix_real

Fiora: fiora \(league of legends\), 1girl

optional tags: streaked hair, white bodysuit, green pants, armor

Headmistress Fiora: headmistress fiora, 1girl

optional tags: pinstripe suit, skirt, red scarf, multicolored hair, hair stick, black thighhighs

Pool Party Fiora: pool party fiora, 1girl

optional tags: striped bikini, red shorts, scarf, ponytail, multicolored hair,

negative embeddings: embedding:bad-hands-5, embedding:ng_deepnegative_v1_75t, embedding:bad-artist, embedding:EasyNegative

Recommended settings:

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

• VAE: vae-ft-mse-840000-ema-pruned (for realistic models & 2,5D), kl-f8-anime2, blessed2.vae (for anime models)

• Sampler: DPM++ SDE Karras or DPM++ 2M Karras

• Steps: 20 to 35

• Clipskip: 2

• Upscaler : img2img in comfyui using 4x-UltraMix_Smooth

• CFG Scale : 5 to 9 (mostly 7 and 8)

• LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.8)

• Denoise strength for img2img: 0.4 to 0.7

If you want the exact same settings, just drag and drop the picture into ComfyUI

for the pictures i used comfyui

LyCORIS of Karen Fukuhara

LyCORIS of Emily Ratajkowski

LyCORIS of Cindy Mello

底模推荐aom3权重0.8、0.7，别的底模也能用自行尝试，后续应该还有一个更新，因为我在炸炉测试中找到一张低loss值跑出的图片很满意但是不稳定，没有特定tag，如果点开例图读取不了tag就随便填，因为我跑例图的这些tag都随机抄的c站其他作者

A LoCon model trained with multires noise & block weights.

This is a LoCon model that trained by 200+ img generated with NijiV5,and it's the multires noise version.

This model can achieve more detailed bright light & shadow which imitation camera lens.

weight 1.0 maybe generates better img.

2023.5.19 Update

InsPX版本不同于之前的1.0和2.0版本LoRA模型，InsPX是全新训练得到的LoHA模型，感觉要比之前的Lora版本效果上要略好一些，建议搭配FilmGirl Provia LoHA 一起使用，否则面部容易崩坏。

请注意，因为这个版本没有触发词，白色画框在某些提示词以及大模型下有一定概率并不会出现，但仍会保持Polaroid 相片的色彩质感。

InsPX LoHA version is different from the previous 1.0 and 2.0 LoRA models. InsPX is a brand new trained LoHA model. It feels slightly better than the previous Lora versions in terms of effect. It is recommended to use with FilmGirl Provia LoHA, otherwise the face is easy to collapse.

Please note that because this version does not have trigger words, the white picture frame may not appear under certain prompt words and large models, but the color texture of Polaroid photos will still be maintained

使用Loha模型的一种方法:

请安装以下插件，https://github.com/KohakuBlueleaf/a1111-sd-webui-locon，然后调用方法与Lora类似，需要将模型放入...\models\Lora文件夹中，随后使用<lora:FilmGX2:XXX>指令调用。

另一种使用Loha模型的方法: → https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris

How to use the Loha model:

Please install the following plugins, https://github.com/KohakuBlueleaf/a1111-sd-webui-locon, then use the same method as Lora. You need to place the model in the \models\Lora folder, and then call it with the lora:FilmGX2:XXX command.

Another way to use the Loha model → https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris

This is a Polaroid-style LoRA, where version 2.0 has become much more stable compared to version 1.0. However, in some complex scenarios, limbs and facial features may still experience distortion. The main reason is that the training set consists of 512x768 photos, and the white frame of the Polaroid occupies a significant portion of the image, which leaves limited content for actual training.

这是一个拍立得风格的LoRA，2.0版本相比1.0版本已经稳定很多，不过在某些复杂的场景下，肢体和面部还是会出现崩坏的情况。主要是因为512*768的训练集照片，拍立得的白色相框占了不小的面积，真正可供训练的照片内容就比较有限了。

If you want to fully replicate the white frame of a Polaroid photo, it is recommended to set the Lora weight between 0.55 and 0.7. Lower weights can display the effect of the Polaroid filter, but the frame size may be inconsistent.

如果要完全复现拍立得照片的白色相框，建议的Lora weight为0.55~0.7。较低的权重可以呈现拍立得滤镜的效果，但相框可能会时大时小。

It is suggested to use this LoRA in combination with FilmGirl LoRA for a synergistic effect in terms of appearance and color correction.

建议搭配FilmGirl LoRA使用，可以起到在形貌和色彩上的增益矫正效果。

After testing, this model works quite well with MoonMIX, majicMIX realistic, and ChilloutMix. Other realistic models should also perform well, so feel free to experiment with different options.

测试了下，本模型在使用MoonMIX、majicMIX realistic和ChilloutMix，效果蛮不错。其他真实系模型应该也不差，大家可以多自行尝试。

Occasionally, the generated photos may show body parts extending beyond the white frame. However, this minor flaw does not detract from the overall image and may even add a sense of depth to the picture.

生成的照片偶尔会出现身体超出白色相框的情况，不过这个感觉无伤大雅，反而会有点立体感增强的效果。

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

LoRA commissions open! -> Fiverr

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

Trigger words for V1.20:

<lora:BreastFeedingv120:1>(breast sucking, mature female, 2girls, yuri, lactation)

<lora:BreastFeedingv120:1>(breast sucking, nursing handjob, mature female, mature male, penis, testicles, lactation)

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

Sample images generated from Nothing model.

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

It was trained on 219 512x512 images scraped from the internet and 20 epochs using animefull_final.

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

Betty Gilpin

LyCORIS of Ashley Johnson

LyCORIS of Alissa White Gluz

LyCORIS of Sylvia Hoeks

This was a requested model. If you want a specific person to appear (or appear sooner) then you can do it by supporting me on my ko-fi page :)

Of course, I will keep adding new people regardless, but if you want someone specific - you can use the ko-fi :)

english↓

This is lora by yukishiro mahiro, a member of the virtual youtuber group nijisanji.

The strength of lora is recommended to be around 0.8~1.

Trigger words:1 girl,yksr

If her hair does not appear, it may be easier to get it if you put "sidetail", "hair ornament", or "blue flower ornament" in the prompt.

It also works well if you put the words that occur frequently by putting lora itself through analysis.

This may not be a lora with satisfactory accuracy since it was made as a trial.

Please enjoy it if you like.

https://www.anycolor.co.jp/guidelines/

japanese↓

これは、バーチャルyoutuberグループのにじさんじに所属している雪城眞尋のloraです。

loraの強度は0.8~1くらいがおすすめです。

トリガーワード:1 girl,yksr

彼女の髪型が出ない時は、プロンプトに「sidetail」「hair ornament」「blue flower ornament」を入れると出やすくなるかもしれません。

また、lora自体を解析にかけて頻出するワードを入れてみると効果があります。

お試しで作ったので、満足のいく精度のloraではないかもしれません。

よければお楽しみください。

Since someone requested it, I will try to create one. It was challenging to find suitable training data. After searching extensively, I found only a dozen or so images. The results are currently not stable, so you need to try several times to achieve the desired outcome.

既然有人要，那我尝试做一个。训练集真不好找，翻遍了才找到十几张，目前效果还不稳定，要多试几次才行。

anythingv4.5,weight around 0.7

This model is my first lycoris model and probably last until i can get a new gpu. It took too long to train and just barely better than lora.Anyway don't use it over 0.7 weight otherwise the image will burn to a crisp :D For base resolution you might need to start above 512x768(I used 768x1152) and hiresfix with 0.2 denoise.You also need lycoris and additional networks extension in order to work this model.Hope you like it :)

I also take a little break from Turkish people.I have a couple foreign people that I want to train on.Don't worry though my Turkish people list looks like a monolith :D

Credits (Because civitai doesn't recognize them even though I installed the extension):

EnvyBetterHands LoCon : https://civitai.com/models/47085?modelVersionId=55199

negative_hand Negative Embedding :https://civitai.com/models/56519/negativehand-negative-embedding

Unspeakable Horrors (Negative prompt) : https://civitai.com/models/4499/unspeakable-horrors-negative-prompt

EasyNegative : https://civitai.com/models/7808/easynegative

Bad Prompt : https://huggingface.co/datasets/Nerfgun3/bad_prompt

Bad Image : https://huggingface.co/khanon/lora-training

培训讯息

第一版本v10训练步数1800步(第二版本V20训练3600步)，v10与v20可两个版本同时使用，基于sd1.5模型训练，请根据使用的模型调整权重。 Clip跳过层1。

训练使用86张高清无码无水印去除背景的NSFW照片,全部皆为裸体照片，手动标记提示词,300张正则化照片。

擅长分开阴部、分开肛门、分开肛门。

培训照片的解析度512x768,768x512.(脸和全身和阴部特写)

请勿再评论发布裸体照片。

这是一个 LyCORIS (LoCon/LoHA) 模型，需要在 Automatic 1111 中额外扩展才能工作。

范例测试使用的模型:

braBeautifulRealistic_brav5

chilloutmix_Ni

realisticVisionV20_20

触发词

触发词:A woman named Anya01,Anya01

额外可用提示词: asian,Detailed skin texture, detailed skin pores, high skin detail,mole,mole_on_breast, nude,completely naked, a woman with a hairy pussy, on all fours, legs apart, simple background, White background, LABIA MINORA,nympha,LABIA MAJORA,everted labia,clitoral hood,urethra,Goosebumps,urethral opening,vaginal opening,

sitting, lying, Moist pussy,small breasts,rough areola,long hair, tattoo

spread pussy,gaping, spread anus,spread ass, puckered anus,indoor, upper body,medium close shot,pussy close-up,face close-up,photo_\(medium\),

photorealistic,Realistic,anal hair, pussy juice,pussy_juice, fingernails,fingers,hands, hairy pussy,excessive pubic hair,mismatched pubic hair, from behind,ass focus,butt lift,m leg,ass in the air, laughing, legs up,from side,

如果你喜欢我的工作并想支持我：https://ko-fi.com/O5O2K9TV1

免责声明：

我不支持公开散布此模型中所有人物生成的任何不雅图像，我不支持出售此模型中人物生成的图像。请不要使用他们在网上发布的不雅内容。

TRAINING INFO

The first version v10 has 1800 training steps (the second version V20 has 3600 training steps), Both v10 and v20 can be used at the same time,based on the sd1.5 model training, please adjust the weight according to the model used. Use Clip skip1.

Training using 86 high-definition uncensored NSFW photos without background removal. All are nude photos. Label the prompt words manually. 300 regularized photos.

Good at spread pussy,spread anus,spread ass.

The resolution of the training photos is 512x768, 768x512. (face & full body&pussy close-up)

Don't comment on posting nude photos anymore.

This is a LyCORIS (LoCon/LoHA) model, and requires an additional extension in Automatic 1111 to work.

Models used for example tests:

braBeautifulRealistic_brav5

chilloutmix_Ni

realisticVisionV20_20

TRIGGER WORD

Trigger Words: A woman named Anya01,Anya01

Additional prompt words available: asian,Detailed skin texture, detailed skin pores, high skin detail,mole,mole_on_breast, nude,completely naked, a woman with a hairy pussy, on all fours, legs apart, simple background, White background, LABIA MINORA,nympha,LABIA MAJORA,everted labia,clitoral hood,urethra,Goosebumps,urethral opening,vaginal opening,

sitting, lying, Moist pussy,small breasts,rough areola,long hair, tattoo

spread pussy,gaping, spread anus,spread ass, puckered anus,indoor, upper body,medium close shot,pussy close-up,face close-up,photo_\(medium\),

photorealistic,Realistic,anal hair, pussy juice,pussy_juice, fingernails,fingers,hands, hairy pussy,excessive pubic hair,mismatched pubic hair, from behind,ass focus,butt lift,m leg,ass in the air, laughing, legs up, from side,

If you love my work and want to support me: https://ko-fi.com/O5O2K9TV1

Disclaimer:

I do not support the public distribution of any indecent images generated by all the characters in this model, and I do not support the sale of images generated by the characters in this model. Please do not use the indecent content they post online.

Yubari | Stylized「LoCon」

Yet another Style LoCon.

Yubari likes to draw a floral print background in most of their art so to avoid overfitting on that detail images with floral background are tagged with floral background

Relaese Version is a Dynamic resize using Sv_fro

🔧 Recommended Usage | Required dependencies

• Loading it natively via LyCORIS extension

❌ Try to avoid

• Loading LyCORIS via the KohyaSS additional Networks extension can result in faulty outputs.

• Vladmandic's fork of Auto1111webui is a poor fork of webui that is bloated with useless features and is only popular because it comes pre-packaged with Torch 2.

My models are made to be used at [ Weight 1 ] Unless it is being combined with other Networks.

📦 Parameters

• Trained on 896x896

• Bucketing 512 x 1024

• 16 Linear dim | 16 linear alpha

• 16 conv dim | 16 conv alpha

📖 License

This collection is released under the CC BY-NC 2.0 license. The license specifies the following terms:

1. You are free to copy, distribute, display and perform the work, and to make derivative works, under the following conditions:

2. You must give the original author credit.

3. You may not use the work for commercial purposes.

4. You may Adapt: remix, transform, and build upon the material.

5. For any reuse or distribution, you must make clear to others the license terms of this work.

   Exceptions : Unless the style of this model is merged enough with different styles making it not easily distinguishable then [3.] does not apply to such derivative models.

   ⚠️ Disclaimer

   The author has no control over user-generated content. Users are responsible for generating NSFW or SFW content. The trained models do not contain obscene visual content that can be viewed with a single click.

Michael, Archangel of Humility from the Seven Heavenly Virtues (Nanatsu no Bitoku).

LoHa trained on NAI. Requires the LyCORIS plugin.

Trigger words:

• michael- For character. Add "red eyes, multicolored eyes" for best results.

• michaeloutfit - For main outfit. Add "uniform, hat, epaulettes, black thighhighs, white skirt, white gloves, black cape, cleavage, wings" for best results.

• michaelidol - For idol outfiit from game. Add "black skirt, black gloves, very long hair" for best results.
  

Now for the Fire Emblem Fates Style LoCON.

This reproduces the portrait art style from Fire Emblem Fates.

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

Mix it with the other Fire Emblem Style LoCONs I have for some fun results.

Keep in mind, this is a style LoCon so it can reproduce character's styles, but not exactly the same character. If you mix this with character LoRAs, then it'll be really accurate.

If you'd like to support me and do more:
https://ko-fi.com/edgonkofi - commissions open! I can make a model for you!

Royal Gowns

We do like to dress to impress, don't we? What's more, we are lacking in beautifully detailed clothing.

Fear not friends. I've got your back.

More beautiful clothing is incoming, beautiful royal gowns to please your eyes and dress your women with style!

Special shout out to @Zovya, I am very impressed with the results of the clothing on her model as I am showcasing it here.

I highly recommend you download either her model A-Zovya RPG Artist Tools - V2

or my Amixx - v1.0-pruned-fp16

when prompting with these models.

Have fun!

Download this LoCon and get some gowns!

今沢 - 艺术家风格 🎨

这位萝拉的画风的原作者是今沢，大家可以去pixiv看看原作者的作品

这是我第一次尝试制作画风lora,实际看起来效果还不错,在更复杂的描述下能取得更好的效果.如果你能喜欢的话,希望你能留下你生成的图片或者评论建议.

使用提示：

• 经常使用的是权重为1到0.7，

• 本次训练和示例图使用模型均为anything-v4.5

• 添加jinze到提示中将强制图像成为艺术家所涵盖的主题，同时赋予 LoRA 更大的影响力。

"Imasawa - Artist Style 🎨

The original creator of this Lora's art style is Imasawa. You can go to Pixiv to check out the original artist's works.

This is my first attempt at creating the Lora art style, and the actual result looks quite good. With more complex descriptions, better effects can be achieved. If you like it, I hope you can leave your generated images or give some comments and suggestions.

Usage Tips:

Frequently used weights range from 1 to 0.7.

The model used in this training and example diagram is anything v4.5

Adding "jinze" to the prompts will force the image to conform to the themes covered by the artist, giving LoRA a greater influence."

Satan, Sin of Wrath from the Seven Mortal Sins anime (Sin: Nanatsu no Taizai).

LoHa trained on NAI. Requires the LyCORIS plugin.

Trigger words:

• satan - For character. Add "red hair, purple eyes" for best results.

• satanoutfit - For main outfit. Add "cape, armor, thighhighs" for best results.

Other trigger words. These concepts were poorly trained and are kind of hit or miss, but can give interesting results:

• satanfury - For Dark Fury outfit from game. Add "red coat, leotard, fishnets, boots, braid" for best results.

• satanaxe - For the polearm/axe. Add "polearm, axe" for best results.

Though the current version might be good at generating the main outfit, it seems rather overtrained. Kind of hard to get the right balance between accuracy and flexibility. I'll probably fine tune and upload a more flexible version in the future. For now, just add outfit elements as negative prompts if attempting to change outfit.

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

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

Upscale it a little to remove any messiness.

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

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

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

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

Recommended weights are 0.65-0.8

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

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

Butterfly Lingerie

Beautiful butterflies...

TL;DR use 0.65 to start and "inverted nipples", inverted_nipples to activate. Increase weight or batch size if desired effect is not achieved.

I'm back! I thought I'd touch up my oldest Lora after learning a few tricks.

This LoRA generates inverted nipples - where the nipple is recessed into the areola. Some are round, some are slanty - we'll get to that.

This LoRA has surprisingly little effect on the visual fidelity of the image, but it's also fragile; some style and character LoRAs are incompatible. Increasing the weight or rerolling the seed can often (but not always) fix this.

I noticed there are quite a few types of inverted nipples while collecting the dataset, with the "slit" pointing every which way. The full LoRA likes to play it safe and generate "volcano"-style nipples in most scenarios. I paired back the dataset to only horizontally-oriented slits for the sideways-only LoRA, which is a little more resilient. I'd use that one, unless you want the volcano-style nipples.

Both new LoRAs are LoCon-Lycoris, so you need the extension.

You can buy me a coffee if you want: https://ko-fi.com/chownyou

Unlimited Blade Works | Weapon LoHA

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

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

This is only a BETA. I will add more Swords, Weapons and Guns to this. SD doesn't like to generate Weapons (just like Ranni with 4 Arms). Trying to improve it on every Update. I recommend using this LoCon for better Hands. If you know how to improve this, i'm happy to hear you out.

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

Trigger Words:

Link Master Sword, Lightsaber, Dual Wielding Swords, One Handed Sword, Unsheathing Sword

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

Weapon Classes:

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

Subcategories:

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

Variations:

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

How to use LoHA's in auto1111:

• Download LoHA

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

• Select your LoHA and use the triggerword

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

• Have Fun!

The New LoCon is much more flexible with a wide variety of checkpoints. Feel free to experiment. The dataset and tagging at the same as the original LoRA. The LoCon remains exceptionally good at generating characters with dark skin.

Oh, I think it's also good for NSFW. I tested it and seems pretty good.

Lastly, you can find some tags from the dataset you can use in your prompts at he bottom.

--Below is for the v2 LoRA
This is a WIP LoRA that replicates the portrait art style of Fire Emblem Three Houses & Fire Emblem Warriors: Three Hopes.

This was trained at 768x768 so it will also best generate images at 768x768 or larger.

I recommend 0.75-0.6 weight for this LoRA. Try different checkpoints, some produce steller results and some produce garbo. Just experiment around, its fun! I use my buddy's Dutch Olympus Model and Dreaming of Olympus which may be released soon if all goes well.

This LoRA can also replicate the different emotions that the character's portraits displayed such as Angry, Blush, Happy, Sad, or Surprised. Just a note. I had better results with Surprised, Open Mouth instead of just Surprised.

This LoRA can replicate the style of the Garrag Mach Monatary outfit as well.

PLEASE leave a review & your comments & opinions because this will not be the final version. I can't even choose between a LoRA or a LoCon.

Also, expect more FE Portrait LoRAs in the future! FE13, FE14, & FE15 are next!

Here are some tags from the dataset you can use:

• simple background

• upper body

• armor

• long hair

• looking at viewer

• closed mouth

• cape

• Young

• garreg mach monastery uniform

• breasts

• uniform

• Older

• green eyes

• Aged Up

• smile

• blonde hair

• open mouth

• Neutral

• Angry

• Happy

• Sad

• Surprised

• Blush

• fur trim

• official alternate costume

A buddy of mine suggested I make a LoCon for the style used in the animated music video, We Are Magonia - The Living Will Envy The Dead Animated Music Video.

The LoCon works great around 0.75, 0.8.

It also mixes with other LoRAs very well.

I strongly recommend you mix this LoCon with https://civitai.com/models/16141/umi-ai-ghosts-and-gods-by-dutchalex for some really cool results. Keep the weight low at 0.2-0.4

Here is a list of tags that you can use in your prompts.

• monochrome

• black background

• solo

• simple background

• red theme

• 1girl

• no humans

• silhouette

• greyscale

• 1boy

• male focus

• skull

• red background

• horror (theme)

• upper body

• holding

• looking at viewer