info -> HuggingFace

A world of 2D illustrations that come to life with an explosion of colors and a magical three-dimensional touch.

Creativity is the stage where imagination and innovation meet, just as rock 'n' roll is the stage where artists channel their passion and energy. Both are powerful means of expression, capable of transmitting intense emotions and touching people deeply. Just as rock music can make us vibrate and stir our emotions, creativity involves us in a whirlwind of ideas, awakening sensations and taking us to new dimensions.

Recommends

Positive Prompt:

(masterpiece, best quality:1.4), (beautiful, perfect, delicate, detailed, intricate, aesthetic:1.2), ...

Negative Prompt:

EasyNegative, BadHands, BadArtist, BadArtistAnime, BadPrompt, BadImages, (low quality, worst quality:1.4), (monochrome, grayscale, sketch:1.2), (logo, signature, text, words, username, artist name, watermark), (loli:1.2), ...

Embeddings (Textual Inversion): EasyNegative, BadHands, BadArtist, BadArtistAnime, BadPrompt, BadImages

VAE: kl-f8-anime2 VAE

Merge Recipe

Merged models are the following:

• Counterfeit-V3.0

• DucHaitenNiji

• NabiMix

Steps:

1. BaseModel = Counterfeit x 0.7 + NabiMix x 0.3

2. MixTape = DucHaitenNiji x 0.3 + BASE_MODEL x 0.7

License

All used models here have "Have different permissions when sharing merges" license permission, so I do not try to add any restriction on it.

CreativeML-Open Rail-M

This model permits users to:

✔Use the model without crediting the creator

✔Sell images they generate

✔Run on services that generate images for money

✔Share merges using this model

✔Sell this model or merges using this model

✔Have different permissions when sharing merges

This is my current most useful Heavy Ink checkpoint merge of iComix, GigerAlienFactory, and the LoRas: FantasyClassics and JimLee Offset.

It is useful for creating extra dark, heavy ink-look illustrations. It might be possible to generate NSFW, but doesn't trend towards responding with that type of content.

Good for generating RPG and storyboarding elements and backgrounds within this heavy handed look.

The included images posted here were all just basic generations with simple prompts to check the response variety. Hard to get it to do front lit, bright cheery picture.

Use your imagination.

Recommendations for results like the first photo:

• Sampling Method: DDIM

• CFG Scale: 30

• Upscaler: SwinIR x4
  

what is "SSSSLLDDLL"

check this prompt, its that all.

shiny sissy luxury latex dress for doll, subtle ambient glow, global illumination, lora:flat2:-1, Negative: EasyNegativeV2,badhandv4,worst quality,(noise:1.1),unnatural pose,improper proportions,facial asymmetry,disconnected elements,irrelevant background, stain, animal ears,

V3 is comming!

This merge model was created with the goal of reflecting the beauty of latex rubber enamel in a anime like illustration; V3 maintains that goal, but with less body shape and a prettier face. i also aimed to create a 2.5D model that is slightly closer to the illustration than before, making it easier to use.

Therefore, SSSSLLDDLLv3 can be used as a normal anime model. As shown in this image, even a very ordinary Japanese adult woman can be output in a cute way, confident in the expression of eyes.

standing, wide angle, (1girls:1.3), harem anime official art, seifuku clothes, school uniform, wide angle, glossy detailed pantyhose, large eyes big, cute beautiful eyes detailed pupils cute, circle white pupil cute, town park background, 35mm f1.4, iso100, raw, high quality,subtle ambient glow, raytracing, global illumination, ultimate quallity,

But you guys who see this image will think: "AYAYA I want to put this woman in a tight-fitting latex suit. No, I bet she has a rubber suit in her bag!" , Absolutely, this model was created to make your dream come true.

i have achieved fabric details like live-action while maintaining the cartoon face. i believe this is "art that is difficult to achieve with illustrations or 3DCG, but can only be expressed with AI models" !

sitting desk, wide angle, cowgirl position, ([precure:lovelive:0.1] face:1.1), (black serafuku:1.2), (fusion black tight school uniform and leotard with sissy frills:1.3), black cameltoe, pantyhose, black gloves, wet clothes latex, lip gloss, makeup, sweat, heavy breathing, (hime cut:0.8) ,black hair, (school classroom:1.2), evil smile, (vulgarity:1.2), school girl, blue eyes, skin glossy, shiny clothes, glistening fabric insanity detailed, subtle ambient glow, global illumination, lora:flat2:-0.4:NF,

BREAKING: v32 "Fixed Face" available now!

2023-05-28 updated!

check "About this Versions" !

Why no V2?

Actually, V2, which was created with a different model than V1, was also partially published, but it suffered from low stability of the limbs. However, when we merged the GOD-LIKE LoRA, both the correction of the human body and the face of the character were improved. im decided to call this V3.

wide angle, missionary, ((kinky harajuku kei)), (sissy:1.4), (silky satin sissy long dress with many ruffles:1.3), (textile shading lolita fasion satin dress:1.3), (Glossy clothes luster:1.3), garter straps, (thighhighs:1.1), gothic latex corset, boots, (torpedo breasts:1.0), heavy makeup, eyeshadow, eyebrows hidden by hair, cheek blush, thick rouge lips, narrow waist, wide hips, large eyes big, circle white pupil anime cute, imperial palace room, subtle ambient glow, global illumination, ultimate quallity

The name of that LoRA is Big Hard Latex. As a side effect, I think the latex outfits are now designed closer to what is actually sold. It is still LATEX ......! LATEX solves everything ......!

Apart from that, we made the busts output with HUGE breasts larger.

((goddess)), (textile shading Navy Blue pvc suit with many ruffles), (angel wing, angel halo), gothic corset pvc, ((torpedo tits, wide hips, gigantic boobs, narrow waist, oppai loli)), Gray hair, seductive smile, in heat, , shiny clothes, glistening fabric insanity detailed, subtle ambient glow, global illumination,

tips (simple or detailed)

If you are trying to produce a simple expression with this model, the following settings are recommended. Also, be sure to use Hires.

Sampler: DPM++ 2M (or DPM++ SDE) CFG scale: 4-6 Hires upscaler: 4x_RealisticRescaler_100000_G (or ESRGAN4x AnimeVideo) Steps:over 16 Hires Steps: 0 or half of Steps Hires Denoise: 0.6

Using this image as a basis, make various changes. Unmodified from below this image (fingers, etc.).

1girl, dynamic pose, dynamic angle, cinematic shadow effect, looking at viewer, cygames mecha_musume, Transformation by mechanization, demon lord girl, (kawaii:1.2), elf ears, demon horns, (metalic bodysuit:1.2), ([Mechanical exoskeleton:boned armor:0.5]:1.0), (glowing mechanical wings:1.2),((floating_armor)), female focus, beautiful detailed eyes, post apocalypse, subtle ambient glow, global illumination, ultimate quallity,

From here, upscaler was changed to latent and CFG was changed to 3. The clarity was reduced by the CFG effect, but the detail was improved. CFG 7-8 is recommended for a clear picture or for a high gloss effect. 9 or higher may result in poor rendering around the lips.

Also, by setting Sampler to UniPC, 3-5 CFG, and setting the Hires Steps to a value less than the Steps value, you can create a 3DCG or figure-like atmosphere. Conversely, the more you increase the value of Hires Steps, the more detailed the writing will be.

I was merging models while looking for other settings that would beautify the gloss, and in the last two weeks or so, two game changers in image representation have emerged. (so often the case in the AI industry LMAO)

One is the extension Self Attention Guidance. This can be done by setting the intensity from 0.75 to 1, which makes the image look like it has increased clarity. This is similar to how depth of field is adjusted with f-stop in photography. In the example, the simple background was left as is, but the background writing may be improved as well.

The other is iroiro-lora's flat2. This LoRA can be applied in the negative to increase detail/writing. In this example, in addition to the above SAG, we have given it a flat2:-1.

These two topics have synergy with each other. Other prompts such as costume wrinkles and hair detailing could be improved, tiled diffusion should also work, etc. All of these are my favorites! Please try them and experience the detail !!

highest quality, (trending on artstation), poster illustration, (2girls:1.2), (a High Quality Victorian Era cute girl), (luxurious intricate-harem dancer outfit with pelvic curtain:1.2), (cameltoe:1.2), anklet, (neck ornament and cleavage and covered nipples), bob cut, (intricate halter backless dress:1.2), sideswept hair, pixie cut, (see-through_dress:1.2), (beautiful face), (sexy eyes), detailed face, detailed eyes, slight smile, glitter many particles, lora:flat2:-1,

[EXTRA] I realized after writing these that CFG1.5 is the practical level only for UniPC. You are not hearing me wrong, it is CFG1.5, use it with SAG and flat2.

misc.

Finally, any liability arising from the use of this model and the media generated by this model is solely that of the user.

We have not tested whether it can be generated for NSFW media. Nudity, etc., ¯\(ツ)/¯ is often detected by AI in each publication, so please be aware of regulations and the laws of your country when using this service.

But even with less skin tone, you and your proclivities can create sexy, and masterpiece artwork.

Only on the LATEX.

standing, wide angle, (1girl:1.0), corruption, (dark persona magical girl:1.2), colorful hair, colorful clothes, dynamic poses, wide angle, latex gloves, highleg playboy bunny suit, showgirl skirt, glossy detailed pantyhose, large eyes big, cute beautiful eyes, effect in dark background, high quality,subtle ambient glow, raytracing, global illumination, ultimate quallity, lora:flat2:-0.6, lora:wowifier_v2:0.5, lora:Villagesuccubus_v30:0.8:NCNF

more infomation and v1 misc, check huggingface.

可愛いと綺麗を併せ持つようマージしたモデルです。

This is a model that combines cuteness and beauty.

使い方(How to use)

・ネガティブプロンプトにはEasyNegativeV2の使用を推奨します。

　I recommend using EasyNegativeV2 for negative prompts.

・VAEはお好みのものをご使用ください。

　Please use your favorite VAE.

・CFG Scale: 4～12

・Sampling method: DPM++ 2M Karras

・Sampling steps: 20～40

・Upscaler: Latent (nearest-exact)

・Hires steps: 20

・Denoising strength:0.5～0.6

上記は使用者の基本設定であり、その範囲外の使用で新たな発見があるかもしれません。

The above settings are the user's basic settings, and there may be new discoveries when used outside of that range.

BeenYou - R12 is up!

For Small breasts & Skinny lovers

R is more realistic, B is basic, more cartoonish

I plan to refine the model with two tracks in the future

⬇Read the info below to get the same quality images🙏

////////////////////////////////////////////////////////////////////////////////////

Recommended Settings - VAE is included

• Clip skip : 2

• Hires. fix : R-ESRGAN 4x+ Anime6B / Upscale by 1.5+ / Hires steps 14 / Denoising strength: 0.4

• Adetailer : face_yolov8n

• Sampler : DPM++ 2M Karras / DPM++ SDE Karras / UniPC

• CFG : 8 Steps : 25

• Prompts

  • (best quality, masterpiece)

  • Neg : (worst quality, low quality)

////////////////////////////////////////////////////////////////////////////////////

Why is my image different from yours?

• Use prompts that match the danbooru tag

  • Tags like Realistic, HDR, etc. have the side effect of forcing the model's characteristics to change

  • Tag auto complete extension

• Use Neg only when necessary

• Refrain from excessive negative embedding

• Use Adetailer for eyes

• Use similar settings

////////////////////////////////////////////////////////////////////////////////////

Do you like my work? check out my profile and see what else!

And A cup of coffee would be nice! 😉

////////////////////////////////////////////////////////////////////////////////////

For anything other than general personal use, please be sure to contact me

You are solely responsible for any legal liability resulting from unethical use of this model

中文/English

所有的预览图没有使用embedding,lora

这个模型的方向是尽可能忠于提示词,能生成更多的东西(人物是主要的,但不代表其他事物的效果不重要),舒服的光影。

你可以用它尝试任何东西！我会不定时补充预览图。如果你觉得这个模型不错,可以评价一下以及上传你的图片。

从AOM3:https://huggingface.co/WarriorMama777/OrangeMixs 和Anything:https://civitai.com/models/9409/or-anything-v5 知道了很多模型和学习了很多混合方面的经验,十分感谢。

使用建议:

脸部出现崩坏的情况,以及想提升面部质量,使用局部重绘 重绘区域使用!!仅蒙版!!(非常建议此方法)获得更好的面部,或使用Hires. fix改善,使用其他随机种或者工具也是不错的办法。

较高的分辨率(比512 * 512高一点)再加上Hires. fix,图片质量会更好(如果显存不够你可以尝试低倍率的Hires. fix或者其他放大方法)。

用于画面质量的正面提示词(像 best quality)是不必要的,会减少画面的可能性,还会使画面趋于一种风格。

将你原本用在正面质量提示词上的权重,用在负面质量提示词上,那是更好的选择。

如果觉得画面内容不够丰富,你可以尝试细致地描述,使画面更加贴近你的想象。

我一般在效果不符合预期时使用clip2

如果想要更强的明暗对比效果,可以尝试光影类lora,比如epi_noiseoffset:https://civitai.com/models/13941/epinoiseoffset

Better light:https://civitai.com/models/33589/better-light

如果你想要更多的亚洲面孔,人物类lora是很好的选择,比如Doll Likeness系列:https://huggingface.co/Kanbara/doll-likeness-series

All preview images do not use embedding,lora

The direction of this model is to be as faithful as possible to the prompt words, to be able to generate more things (the character is the main, but does not mean that the effect of other things is not important), comfortable light and shadow.

You can try anything with it! I will add preview images from time to time. If you think this model is good, comment on it and upload your pictures.

I know a lot of models and learn a lot of mixing experience from AOM3:https://huggingface.co/WarriorMama777/OrangeMixs and Anything:https://civitai.com/models/9409/or-anything-v5, thank you very much

Suggestions for use:

If the face appears to be falling apart, and you want to improve the quality of the face, use Inpaint and Inpaint area use!!only Masked!! (This method is highly recommended) to get a better face, or use Hires. fix to improve, use other seed or tools is also a good way.

Higher resolution (a little higher than 512 * 512) plus Hires. fix, picture quality will be better (if the gpu memory is not enough you can try a Low magnification of Hires. fix or other upscale tools)

Positive prompt for image quality (like best quality) are unnecessary and reduce the possibilities of the picture, also make the picture tend to be in a style.

It's better to Use the weight you would have used for positive quality prompt on negative quality prompt.

If you feel that the content of the picture is not rich enough, You can try to describe in detail to make the picture more closely to your imagination.

I usually use clip2 when the results don't meet expectations

If you want a stronger contrast between light and dark, you can try light and shadow lora, such as

epi_noiseoffset:https://civitai.com/models/13941/epinoiseoffset

Better light:https://civitai.com/models/33589/better-light

If you want more Asian faces, the Character class lora is a good choice, such as the Doll Likeness series: https://huggingface.co/Kanbara/doll-likeness-series

你可以从版本信息中知道当前版本信息,比如使用了哪些模型/You can know the current version information from the version information, such as which models are used

我使用这两个VAE/I use these two VAEs:

https://huggingface.co/stabilityai/sd-vae-ft-mse-original

https://civitai.com/models/22354/clearvae

配方在/Recipe in → huggingface:https://huggingface.co/Fre2C/DreaMirror-Mix

This merge likes simple (but creative) prompts to create art with the creative versatility of the original XenoEngine 4.5 with a huge push toward semi-realism and fantasy. Though made with portraiture and consistency in mind, it does a nice job as a flexible base model for just about anything I've tried to make with it.

Certain prompts tend to go more "art" than "photography" style. Using "Real Life" along with traditional "photography" cues in the prompt can help bring more semi-realism.

The sample images I have made are done with simple prompts and Easy Negative and Bad Prompts Version 2 - no postwork or cleanup. As always - to get really great stuff with SD, inpainting and cleanup is likely needed.

This is the merge model for my personal taste.

Based on my beloved model "Fidelity", I've merged some of my favorite models.

*Basically, I prefer POV images, so I'm tuning them specifically for that. (However, my skills are lacking, and I feel that things are not going well...)

*Recommended prompts, please check the data in the image I'm posting.

*This model is also somewhat suitable for NSFW.

Thanks to the predecessors who have released many models!

I don't know what's inside MIX
Cause I forgot what models was used at merging

THIS MODEL IS NSFW MODEL

Even if you don't use NSFW tags inside your prompt, this model can print NSFW arts.

Recommended cfg scale for 1.0 is 5 and for V2 is 8.

V2 is improved version with merging Abyss Orange Mix 3
V2 is showing little difference on details

Deformed version added!

Merge recipe :
idkwi_V2 + Nukunuku nigirimeshi lora (0.3) + Pchan style lora (0.3)


Here's comparison image

Works perfectly only with promps with zoom in at the face, focus on the face

FaeTality

Please read the description below for model information!

This is the second model mix that I have publicly released. Originally started as a version upgrade for FaeTastic...I fell in love with it and decided that it needed its own unique name and space.

Like FaeTastic this is a very versatile model that does have the Noise Offset included. I still love FaeTastic and I use it regularly in my work still, but FaeTality I find it to be a bit more creative and also makes some amazing eye details.

I use the F8Anime2 VAE for this model, but the 840k also works perfectly well too. If your colors are appearing washed out, then you need to install and use the VAE properly.

Almost all images utilized Loopback Scaler.

I very highly recommend this script as it does some awesome stuff (why the images example images are highly detailed with awesome vibrant coloring!!). For any images not utilizing it, I instead used the SD Upscale. I've written a guide here going over it for those not versed in upscaling: https://ko-fi.com/post/The-Double-Headed-Issue-E1E0JO2UV

When I have time I plan on publishing some article guides on Civit going over the Loopback Scaler and how to utilize it properly.

This model mix has been a several month long process in the making as I am highly picky about the quality of things that I want to release. That's why I don't churn out things on here (I have A LOT of models and TIs/LORAs but am very anxious about releasing them unless I feel like they've attained high standards!). It probably took hundreds of merges and a crapton of playing with merge sliders to figure out things. If you feel like supporting me here is my Kofi.

Any questions related to the model or any help needed with Civitai I am always available to be contacted via Civitai discord. I don't bite. :)

No NSFW experiments, i don't even know what's possible or not possible

Introducing BUN MIX, a bun that is nothing short of spectacular.

BUN MIX is a model in the fantasy model. It was created with the goal of giving you maximum results with minimal prompts.

has a pretty high finger batting average.

I merged this model with the idea of not only hiring but also upscaling it further, so the use of additional upscalers like Tiled VAE or sd upsacle is almost mandatory, as is the use of Dynamic Thresholding.

I won't bother with further explanations or comparisons to other models.

You can use the same option you've been using, and I recommend using a higher sampling step. At least 40 or more. The models I merged use UNIPC by default, so you only need to pay attention to these two things

my DT option ⬇️

Using my VAE will make it feel even best -> https://civitai.com/models/65728/cute-vae

visit me on Twitter -> https://twitter.com/nzxt_cam

If you'd like to help ease the burden of my electricity bill, buy me a coffee! My electricity bill renewed for $100 this month -> https://ko-fi.com/yuribun

This is a global repository of the Dreambooth Checkpoints.

Some people asked me to upload the source Dreambooths from which I make the LyCORIS.

To not spam the site/profile with multiple models I decided to do it the same way I intended to handle the Game characters -> a single page where each model is just a separate version.

Upcoming: Ronda Rousey, Heidi Klum, Cate Blanchett, Betty Gilpin, Lake Bell

(if there are no more suggestions from the community, I will be picking myself who to upload next :P)

NOTE1: I will be uploading up to one Dreambooth per day (the visibility of other models suffer and I want all to have an equal chance :P, also this gives me more time for LyCORIS)

NOTE2: Feel free to suggest who should be added (you can look at my existing LORA/LyCORIS library to see which Dreambooth I have; but of course a completely new person can be suggested as well - however that would be best to do via my ko-fi page unless you can wait very very long :P)

NOTE3: This is for those who prefer to get the maximum quality! If you are fine with the quality of the LyCORIS, you may want to skip it (however, I imagine I might upload some models here before I convert them into LyCORIS for time reasons).

However, If you prefer LyCORIS - you need not download anything from here because I eventually will upload all those models in the LyCORIS version :)

Got a potato for a PC? Use this model for free on Happy Accidents or on the Stable Horde

Updated - v8: Should have the lora compatibility of v6, the details of v7, with a more realistic and less saturated look. See ‘About This Version’ for more details. This photo is the calibration prompt meaning that you should be able to run it on any UI and get the same result due to a short prompt with no embeddings/hires fix/other fancy tweaks

Following on from Gorilla With A Brick, I've merged in 10 more photorealistic models at various weights, and some more noise offset to create something that when prompted for photorealism will make you go "I Can't Believe It's Not Photography". It will happily create CGI characters and awesome landscapes as well.

As always, pruned to fp16, and the VAE baked in (SD-v2 840000)

このモデルは鉛筆のような輪郭線と水彩を少し混ぜたような塗りのイラストを作ることができるモデルです。

This model is capable of creating illustrations with pencil-like outlines and a bit of watercolor mixed in with the paint.

このモデルはcounterfaitV3をベースに2つの自作LoRAを含む三つのLoRAをマージして作成されました。

This model was created by merging three LoRAs, including two home-made LoRAs, based on counterfaitV3.

推奨設定は以下の通りです。

推奨VAE：clearVAE

サンプラー（Sampler）：DPM++ 2M Karras

ステップ（Steps）：20

スケール（CFG scale）：7~9

アップスケーラー（Hires upscaler）：none

高解像度ステップ数（Hires steps）：15

ノイズ除去（Denoising strength）：0.5

The recommended settings are as follows

Sampler: DPM++ 2M Karras

Steps: 20

Scale (CFG scale): 7~9

Hires upscaler: none

Hires steps (Hires steps): 15

Denoising strength: 0.5

If you want to support my work and keep me motivated to produce more content, you could buy me a coffee or tea:

https://ko-fi.com/gswan

https://www.buymeacoffee.com/gswan
------------------------------------------------------------------------------------------
This is the first trained model for Basil Japan..

2023.5.19

更新了Utopia版本。这个版本基于Fantasy2.0模型微调而来。在减少模型体积同时，图像真实感进一步提升。

Updated Utopia. This version is based on the Fantasy2.0 version, fine-tuned. While reducing the size of the model, the image realism is further improved.

2023.5.17

更新了reality 3.0。这个版本在减少模型体积同时，图像真实感与面容美感进一步提升。

Updated reality 3.0. In this version, while reducing the size of the model, the realism of the image and the beauty of the face are further improved.

2023.4.21

我将胶片风Lora小比例融合进了Fantasy模型中，优化了此模型下皮肤的质感。但请注意，要获得完美的胶片质感，仍需要搭配胶片风Lora使用，只是使用比例可以降低0.1~0.15左右。

Fantasy 2.0模型更偏幻想风格一些，Reality 2.0模型更偏写实风格一些。两个模型质量都不错，大家可以根据需求下载使用。

2023.4.20

更新了reality2.0版本，这个版本相比之前两个版本，较大幅度提升了皮肤和样貌的真实感。建议在生成人像时，搭配胶片感Lora（1girlmix）使用，真实感更佳。

同时最重要的，新模型搭配兔狲Lora使用，生成兔狲逼真图像的成功率大幅提升！（没错，我搞这个模型，其实主要是为了训练猫猫）

chilloutmix模型与本模型的reality版本、fantasy版本的对比：

Comparison between the chilloutmix model and the reality and fantasy versions of this model:

我比较推荐用reality版本的模型，在色调、颜值和真实性上比较平衡。建议将clip skip设为1，这样生成的图会更美观。

I recommend using the reality version of the model, which is more balanced in tone, appearance and authenticity.

本模型由多个checkpoint模型合并而来，具体有哪些我也记不太清了。以下是MoonMix与chilloutMix模型在生成人像时的一些对比。总的来说，MoonMix相比chilloutMix，真实感和锐度有所下降，但美感与丰富度有所提升。

This model is the result of merging several checkpoint models. I can't remember exactly which models there are. Here are some comparisons between MoonMix and chilloutMix models in generating portraits. In general, MoonMix is less realistic and sharper than chilloutMix, but the overall beauty and richness is improved.

演示样例中所使用Lora模型“1girlmix”可以在此处获取：https://civitai.com/models/33208/lora

The Lora model "1girlmix" used in the demo example is available here: https://civitai.com/models/33208/lora

拿counterfeit2.5, perfect world, cetusmix, mixprov4, hassakuHentaiModel_v12等等做mbw merging整出来的，就乱塞

Detail Tweaker LoRA (细节调整LoRA) - v1.0 | Stable Diffusion LoRA | Civitai

这LoRA必须用，一般 1 - 1.75左右

negative prompt用咱的sb长prompt，效果可以的

EasyNegative - EasyNegative | Stable Diffusion Textual Inversion | Civitai

Deep Negative V1.x - V1 75T | Stable Diffusion Textual Inversion | Civitai

veryBadImageNegative - veryBadImageNegative_v1.3 | Stable Diffusion Textual Inversion | Civitai

negative_hand Negative Embedding - negative_hand | Stable Diffusion Textual Inversion | Civitai

bad_prompt Negative Embedding - bad_prompt_version2 | Stable Diffusion Textual Inversion | Civitai

Async's Negative - v6 | Stable Diffusion Textual Inversion | Civitai

σ`∀´)全是给大乃吸引过来是吧ᕕ( ᐛ )ᕗ爱看就多整点

3IV是anime模

3VI是anime+semi-realistic模，就是咱mbw塞太多chilloutmix了，有时太韩国女星脸

3VII有两个版本：I 和 II

I是咱3VII(没上传的版本)加了点majicmix进unet细节層

II是借I经验，加了点bloodsuga大佬的模和majicmix，最后把anythingv5的text encoder融了九成进去〔anythingv5 text encoder听说很好使〕

为什么在前面放3VII-II？这样才有人下啊(　ﾟ 3ﾟ)

3VII-III is good too

Cinema_4Dream is the best I can do without training a model on Cinema 4D. Im not saying this is accurate but it’s REALLY FUN!!!!  Show me what you got. 

21 Ingredients…  (First 5 are the spicy triggers): 

1. (FLUIDART: ) 

2. (LIQUID SPLASH: )—FeverDream

3. (Knollingcase: )

4. (FOOD_CRIT: ) —Jak’s Creepy Critters

5. (Beeple:)

6. Color Fusion

7. 3D Magic Space

8. Crystal Clear

9. Holographic Model

10. Unreal Engine 5

11. Slimy’s Super Model V2

12. Babes V2

13. AETHER

14. Artius

15. Orusium

16. Vodka V2

17. AW PORTRAIT

18. Project Cryengine

19. Project Unity Beta V2

20. Dreamful V3

21. RmadaArt V8 Cold Vae

22. Sentinel Diffusion

lovry

AI Avatar photos and model beautiful and women and girls

vae: 840000

• Combining multiple loras trained by the base model counterfleit

• gives a great brush painting style,

• which can be used with other loras.

• If you like my model, please give feedback below or follow me, I will keep updating

• Q群692574838

This model is a result of multiple mixes and lora merges to get cute faces with big round eyes, high contrast, thick black lines with a high level of detail.

The pros :
- Lora stacking without too much mess
- Good with both SFW and NSFW
- Very little prompting required
- Highly detailed skin & clothing textures
- More than decent hands from base gen

The cons :
- Might give some unwanted animal ears
- Do no mix well with certain style lora

Recommended parameters :

VAE : kl-f8-anime2.pt

Masterpiece, best quality

(worst quality, low quality:1.4)

Sampler : Dpm++ 2M Karras Steps : 30

Sampler : Dpm++ SDE Karras Steps : 25

Sampler : UniPC Steps : 30

CFG scale : 6.5-7.5

Clip Skip : 2

Hires.fix: Ultra_Remacri (See Training data) or SwinIR_4X Denoising strength : 0.45

I'd highly recommend to also use Adetailer at default parameters with the "face_yolov8s" model.

Tips :
- For better quality eyes and iris, simply specify eye colors
- Keep the negative simple, stacking negative TI will have the opposite effect

This model is made to achieve the skin detail seen from my initial model Ikigai 2.5D paired with the Danni 2D lora that's no longer available, without any lora needed!

My usual settings:
Sampler: DPM++ 2M Karras
Sample Steps: 20-35
Upscaler: ESRGAN_4x
Deniose: 0.5
CFG: 5-9

By no means is this law, feel free to experiment and I'd love to see your results!!

For inpainting, I generally limit to Only Masked to 800 x 800 or lower and 0.5 denoise (or lower if you want to maintain the initial face look more). It will start to get a little too detailed at higher res'. You shouldn't need to change your prompt at all when inpainting faces to get a good result. This was a goal of mine, as I prefer not to have to do any extra steps in my workflow if possible.

This mix comprises of:
MajicMix Realistic
MIX-Pro-V4
Babes 2.0
FantasticMix

Same as the 2.5D model, this one plays very well with 2D artist style loras!

brand new anime mix model,3d version.

The prompt "fisheyes" is very important ,or you cannot output the eyes like the Cover image.

An anime model with a sharp, colorful style. Has KL-F8-anime2 VAE baked.

It doesn't take much to get a good-looking image, none of the preview images even use a negative embedding.

If you like my work, consider supporting me: https://ko-fi.com/mira6

1girl sit with smile, rise leg, outdoor, the most beautiful image ever seen, uhd, RAW, sharp focus, (photorealistic), best quality, ultra high res, masterpiece <lora:koreaface10:0.8>

Negative prompt: (3 legs), (3 arms:1.1), (worst quality:2), (low quality:2), (normal quality:2), lowres, acnes, skin blemishes, age spot, glans,extra fingers,fewer fingers,strange fingers, ((monochrome)), ((grayscale)),(bad legs:1.5), (bad hands:1.5), (3_hands:1.5),

Size: 512x768, Seed: 1582028139, Model: chilloutmix_Ni, Steps: 35, Version: v1.2.1, Sampler: DPM++ SDE Karras, CFG scale: 7, Model hash: 7234b76e42, Hires steps: 25, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+, Face restoration: CodeFormer, Denoising strength: 0.5

Insane Isometric

Original Finetune based on https://civitai.com/models/20562

This is largely an ISOMETRIC model, but it can still do ANIME babes, and stuff - it's just that it's predominately a style we've been wanting to re-do so we can mash together our old Isometric models!

KEYWORDS THAT ARE LITERALLY JUST TO GET THE TRAINED STYLE:
ISOMETRIC OR NO HUMANS - it is up to you.

Also DONT DO ANYTHING ILLEGAL WITH OUR MODELS or i'm gonna cry, and blame it all on you.

---

Join our Reddit: https://www.reddit.com/r/earthndusk/

Funding for a HUGE ART PROJECT THIS YEAR: https://www.buymeacoffee.com/duskfallxcrew / any chance you can spare a coffee or three? https://ko-fi.com/DUSKFALLcrew

If you got requests, or concerns, We're still looking for beta testers: JOIN THE DISCORD AND DEMAND THINGS OF US: https://discord.gg/Da7s8d3KJ7

Listen to the music that we've made that goes with our art: https://open.spotify.com/playlist/00R8x00YktB4u541imdSSf?si=b60d209385a74b38

---

license: creativeml-openrail-m

language:

- en

tags:

- finetune

- isometric

- landscape

- anime

---

### Insane_Isometric Dreambooth model trained by Duskfallcrew with TheLastBen's fast-DreamBooth notebook

### THIS CAN DO MORE THAN ISOMETRIC - but it's largely MEANT for Isometric and Landscapes,

### To get mostly NOT humans use: ISOMETRIC OR NON HUMANS Or both!

### WORKS ON DANBOORU TAGS!

GEN SERVICES PLEASE CREDIT, AND IF YOU MAKE MONEY THROW US SOME CAUSE WE BROKE.

日本語の説明は後半にあります。

This is a merge model for generating bright-toned doll-like / 3D-animation-film-like illustrations.

Although this model is doll-like, it is not as if the doll were placed as is, but rather as if only the texture of the 2.5D illustration, including the background, were transformed to be doll-like. A block-merge with multiple doll-like models is utilized with an awareness of balance as well as stable response to prompts.

The main models merged are as follows.

• foxyaNSFW v3, DucHaitenNiji v2.35, AOAOKO PVC Style, FaceBombMix v1, etc.

Both SFW and NSFW outputs are capable as NSFW models are included in the merge. High quality and unique images can be output in a short prompt. VAE is not needed, but you can try some.

List of my models

• CityEdgeMix : Simple photo-realistic model based on Chilled family

• CityEdge_ToonMix : Simple 2.5D anime-like model based on CityEdgeMix

• CityEdge_2dToonMix : Pure 2D anime-like model merging only 2D models

• CityEdge_StyleMix : Anime-like model with detail depiction by making use of block-merge

• CityEdge_FusionMix : 2.5D model focusing on unique depiction through merging many models and block-merge.

• Landscape_Photoreal : Model that produces photorealistic output, whether with characters or just the background

• CityEdge_DollyMix : Model that produces doll-like and 3D animated movie-like output

明るいトーンのドール調・3Dアニメ映画調のイラストを生成するためのマージモデルです。

ドール調ではあるものの、人形がそのまま置いてあるようなイメージではなく、背景まで含めて描写された2.5Dイラストの質感だけをドール調にしたようなバランスを意識し、複数のドール調のモデルを階層マージしています。プロンプトへの安定した反応も意識しました。

マージしている主なモデルは以下のとおり。

foxyaNSFW v3, DucHaitenNiji v2.35, AOAOKO PVC Style, FaceBombMix v1

NSFWモデルをマージに含めているため、エロ・非エロどちらでも可。VAEは焼き込み済ですが別のものも試してみていいと思います。

モデル一覧

• CityEdgeMix : Chilled系をベースにしたシンプルなフォトリアル系マージモデル

• CityEdge_ToonMix : CityEdgeMixを2.5D化したシンプルな2.5Dアニメモデル

• CityEdge_2dToonMix : 2Dモデルだけをマージしたピュア2Dアニメモデル

• CityEdge_StyleMix : 階層マージを駆使してディテール描写にこだわったアニメモデル

• CityEdge_FusionMix : 多数のモデルのマージと階層マージによる独自の描写にこだわった2.5Dモデル

• Landscape_Photoreal : キャラクターがいても、背景だけでも、フォトリアルな出力が得られるモデル

• CityEdge_DollyMix : ドール調・3Dアニメ映画調の出力が得られるモデル

AnyOrangeMix

AnyOrangeMix is merge model of Anything v4.5 and AbyssOrangeMix3.

Now adding Anything v5.

HuggingFace

Merge Source:

Anything v4.5 (0.5) + AbyssOrangeMix3 A1B (0.5) Weighted Sum

(Anything v4.5 + Anything v5 Prt 0.6 Weighted Sum) +

(AbyssOrangeMix3 A1B + AbyssOrangeMix3 A3 0.5 Weighted Sum) 0.5 Weighted Sum

Merged using my model merger.

Recommended Settings:

• Sampler: “DPM++ SDE Karras” recommended.

• Steps: 20~

• Clipskip: 1 or 2

• CFG Scale: 7 or higher recommended.

• VAE: anything_v4.0.vae.pt baked

Recommended Prompt:

Prompt : masterpiece, best quality,

Negative : lowres,bad anatomy,bad hands,text,error,missing fingers,extra digit,fewer digits,cropped,worst quality,low quality,normal quality,jpeg artifacts,blurry,extra legs,extra feet,extra arms,extra fingers,missing legs,missing arms,ugly ,huge breasts,monochrome

Recommended Embeds:

• bad prompt

• bad hands

• bad artist

• Easy Negative

• Deep Negative

A western-style merge, focusing on bold lines, vibrant colors, and a large variety of themes.

ふわふわで可愛い女の子のイラストを生成するのが得意なモデルです。

The model is good at generating illustrations of fuwafuwa and cute girls.

This model is based on:

Anything V3

黑猫盛宴 | BlackCat RAT

Crowbox-Vol.1

Nostalgia-clear

SomethingV2.1

DARKEST

My Fans Lora : fanslora.com/miaou

prompts exemple/helper : here

Trained with +2k Pictures

Have fun and don't hesitate to share your images ! 

Credite if you merge this model.

A custom checkpoint mix intended to unite realistic people and breathtaking landscapes, with an emphasis on rich colors and lighting composition, and with a healthy dose of sci-fi/fantasy.

No baked-in VAE. I personally recommend EMA 560k, but MSE works well too.

🥷 Ninja V1 by Chillpixel

Get hyped for an unforgettable ride through the trendiest Checkpoints and LoRAs of the moment, all jam-packed into a legendary model that's about to blow your mind.

👋 Hey there! I'm Chillpixel, a developer, just diving into the vibrant CivitAI community.

Ninja aims to be the "Greatest Hits" album of CivitAI; bringing together all the absolute best and most popular community resources in one awesome checkpoint.

This epic model is your ticket to exploring the hottest trends and unleashing your creativity like never before. Get ready to level up and dive into the coolest journey yet!

The example images were made using a custom Stable Diffusion pipeline called Ninja Diffusion, available exclusively at DreamWalker. Select Ninja from the model dropdown list — and crank the steps/guidance all the way up — the results are wild.

I really hope you dig my first drop — and if you do, please hit me up with a 5-star review for Ninja!

Companion LoRA

Introducing Elixir, my enhancer LoRA, which helps revitalize antiquated models, infusing them with renewed vitality. While for newer models, it offers a gentle yet discernible enhancement.

Let's Have Fun!

I'm kicking off daily art themes and challenges now:

👉 https://discord.gg/89Pu5ehUvE

This model was trained from about 60 pictures of the character Yu Lexuan, from "In a class of her own" 漂亮书生 Chinese drama. This is my first model and may have some problems, but the test pictures are quite good.

Credits to Shivam Shrirao for making a great tutorial on how to train models. You can also make your own here using the provided Google Collab notebook.

Beware that this model is unfiltered and will produce NSFW content if asked so.

A cartoon-style template still in the process of being created

"Clam" is named from using modes' name.

It has better performance in txt2img when using positive prompts to describe what you want without any quality tags, and you can use EasyNegative as negative prompt.

Ver.1.1 has more details (like background,body...).

Ver.1.2 has more anime-styled hands (but it may generate strange figures).

无需添加正面质量标签，直接在正面标签里面输入想要的内容，使用EasyNegative或者其它嵌入式负面标签作为负面标签即可出较好的图。

建议使用vae，模型由于通过数次融合已经呈现出比较灰的颜色效果，vae修正可以得到比较好的结果。使用NovalAI官方原版vae（animevae.pt f458b5c6）或者mse-840000等都能呈现出比较不错的颜色效果。

本模型是融合模型得到的平涂风格大模型，CLAM来自最初合并的大模型的首字母或者其中关键字母。后续又不断融合了更多的模型，但是没有继续改名。该模型名称可以翻译为“蚌”。

该模型目的是在保留平涂风的情况下，增加更多有效的细节表现，并克服一些过于写实的表现，并旨在增加细节的稳定性。

1.2版本修复了手的细节，相较1.1更加动漫化，不过调参不好或者在潜变量放大的时候，有一定几率出现一些奇怪的身体姿势。基本很好保留了1.1版本的各种优点。不过有些缺点仍然会比较明显，比如配色没有那么好，用色偏平淡，有种雾蒙蒙的色彩感。光线表现太过丰富，但是又有点不太合理。

1.3版本修复了一些过分的细节感，较好地优化了配色和1.2版本奇怪的身体姿势，在光线表现上也趋于更好。手部的表现与1.2版本相当，不过在手被遮挡或者双手交叉或者位于图片边缘的时候仍然有奇怪的表现，容易和其他东西发生融合。不过一定程度上提高了出图质量，在不使用lora和controlnet的情况下，相比上个版本更容易在文生图的情况下生成违和感较低的图片。建议在高清修复前的生成图像像素长宽控制在1024以内，推荐512*768或者600*960，横图竖图都能出，效果都不错。

有好的意见或建议也欢迎反馈。会在以后版本好好努力的！

hf:https://huggingface.co/Johnhex/Clam

https://huggingface.co/Johnhex/Clam1.1

https://huggingface.co/Johnhex/Clam1.2

https://huggingface.co/Johnhex/Clam1.3

How to Use

Here is the parameter recommended to use in stable diffusion.

VAE: animevae.pt [f458b5c6]

Sampler: DPM++ 2M Karras or DPM++ 2S a Karras

Step: 15~30

Hires Upscale: 2,Latent Denoising:0.5~0.8

Hires Upscale: 2,R-ESRGAN 4x+ Anime6B Denoising:0.2~0.6

CFG Scale: 7~15

Edit: I saw that we forgot to update the Modelversion in the Metadata. But don´t worry, that output is still from 1.9 . I was simply too tired to make them again ;)

Update 1.9

Chillpixel's NinjaFix takes Stable Diffusion models to new heights, producing stunning visual art through precise neural network fine-tuning, improved accuracy, and stable output.

With Juggernaut 1.9, we are paving the way for Version 2. (Release hopefully in 2 Weeks)

The base model has been improved.

NinjaFix has also received a brand new update for Version 1.9 and is even better than before.

If you want the NinjaFix with HiResFix (True Enhancer) on Juggernaut , then join DREAMWALKER . I test all my models there with the Community, so if your PC Hardware isnt good enough (like mine :D ) then take a look at Dreamwalker :) Of course there are other models too ;)

The Examples are made with the True Enhancer from Dreamwalker. But of course you can also use other Upscale Versions

I would appreciate a review. And I'm also looking forward to seeing your images using this model :)

By the way, you can follow me on Instagram

Or you can buy me a Coffee/Beer (but it's really not necessary, I would still appreciate it though :D)

Today, I'm presenting to you "Juggernaut," a side project that emerged during my work. In the past, I was always annoyed when I had to switch models to achieve different styles, so I had been working on another model (StaVis) for some time. From that, this side project emerged, which didn't fit my original vision but was still too good to be thrown away.

Juggernaut is an extremely versatile model that can serve several style directions (paintings, digital, realism, anime, NSFW) and delivers absolutely sharp quality. The sample images are the raw output from Stable Diffusion. The standard VAE was used, so you can experiment with your own VAEs.

It started a couple of weeks ago when i started training my own model (with round about 5k images (50 % Hand tagged/50 % AI Tagged). It was still not the model i searched for so I merged a couple of models into this (Dont ask me which models, i simple didnt write it down) . This Model is more "NSFW" as the Original Project (StaVis) and prob will be updated more into nsfw in the future. But in the End you guys decide ;)

The model works very well with weighted prompts but has some difficulties with shorter prompts.

I hope you have as much fun with "Juggernaut" as I do :) I would be happy to see some of your images in the gallery below.

DOWNLOAD NEGATIVE EMBEDDING HERE FOR BEST PERFORMANCE.

This is Aether Lux, a Lexica-inspired 2.1 model, trained on synthetic data. Images are preferably generated with a base side of 768px, and for general purpose, Euler works well. For photorealism: DPM++ SDE Karras is the best choice in my experience.

The model is accompanied with a negative embedding, lux_np. You need this one applied to the negative prompt to make sure that the model functions at its very best. You place the file in the Embeddings folder within your Stable Diffusion installation (usually stable-diffusion-webui\embeddings).

For this model, I have set out to depict typical anime appearances of the newer generation with a touch of realism, without compromising the character's style. It was also a priority for me to ensure a diverse range of colors.

Furthermore, this model is well-suited for depicting explicit content.

__________________________________________________________________________________________________

Recommended parameters:

((masterpiece)), (best quality),

Negatives: (worst quality, low quality:1.4),

EasyNegative, bad-hands-5,

Sampler: Euler a: 40 to 60 steps.

Sampler: DPM++ SDE Karras: 30 to 60 steps.

Sampler: DPM++ 2M Karras: 20 to 60 steps.

Sampler: DPM++ 2M SDE Karras: 20 to 60 steps.

CFG Scale: 4 to 11.

Resolutions: 512x768, 512x1024 for Portrait!

Resolutions: 768x512, 1024x512, 1536x512 for Landscape!

Hires.fix: Latent (nearst-exact), with 20 steps at 0.54 - 0.7 denoising, upscale 1.85-2.0.

Clip Skip: 2.

__________________________________________________________________________________________________

Bad-Hand-5

https://huggingface.co/yesyeahvh/bad-hands-5/tree/main

__________________________________________________________________________________________________

Base of this Model is MeinaMix v10

other parts of my model

hassakuHentaiModel_v12
abyssorangemix3AOM3_aom3a1b
AnythingV5_v5PrtRE
BloodorangemixHardcore_bloodorangemix
animePastelDream_softBakedVae
breakdomain_M2150
koji_v2

超不喜欢目前的日漫风格，国漫也不喜欢，相比之下早期日漫还凑合，所以合成一种早期漫画风格，尽量远离写实和2.5D，试图与ANIME FINAL融合，后者简直了...

所以a+b搭配如下：

Darksoul:0.5

Nightsky:0.5

oldfish:0.75

chinese weding:0,5

This model<2.5D,hope so.

Hi there, and welcome to my first ever checkpoint!

What I decided to do with this is create an illusionary combination of Western cartoon and anime styles, and also add 3D photorealism to the mix. It's something that I personally like to call:

CrossoverMix.

Recipe (and I credit everyone who made these models):

0.55(0.6(toonyouB31.MGIm) + 0.4(AnythingV5)) + 0.45(realisticVisionV20_v20) + 0.7(revAnimated_v122 - lyriel_v15)

Recommended samplers: Euler A, or DPM++ 2M Karras

Clip skip: 2

CFG Scale: 5-7

Upscale by: 2

Denoising strength: (if you're doing hires. fix, do ~0.55-0.7)

I didn't bake any VAE in, but I highly recommend these VAEs:

vae-ft-mse-840000-ema-pruned

Night Sky VAE (located at the download link)

Suggestions and feedback are welcome! Hope you enjoy!

Watercolor style background ckpt, can draw people but not many

水彩背景特化模型，能画人但不多

It's a merge of models as well as my own trained models and LoRas merged together to make this. Makes furry content primarily but can do human as well. It's basically just a mashup of a bunch of stuff.

a blend/merge of many of my models , along with some of yours too!

Vae baked in.

fp16

clip skip 2 recommended but will work iwith any.

It is a model that specializes in expressions between 2D and 3D.

It is recommended to use a sentence like (line drawing extraction:0.5).

The broken hand is not severe, It is easy to generate neat pictures without using "Hires.fix".

its mixing the following

MagMix , BreakDomainAnime , Defacta 5rd , Noosphere

> VERSION 2.5 NOW AVAILABLE! FULL & PRUNED VERSIONS <

📜Many years ago... when I got first got my start in artwork.. some of my first attempts were to draw comic book characters. Over time during my career I've had the fortune to work on graphic novels and even with some pretty big names. There is an old love that runs in me of this genre. This is my attempt to capture that and bring it to SD!
===========================================================

☞(Please read below the Pro Tips & Tricks section for the vae & necessary settings to get the best results!)☜

A hand selected collection of the best of the best, no compromise models to make up this collection!

Specializes in traditional Western Style Comic Characters 2D-2.5D Semi-Realistic, Illustrations & Fantasy Art!

Will do everything from sketch designs, to full comic posters, character concepts, panel layouts, complex scenery, fantasy art of all kinds, sci-fi, retro, etc. The results are endless!

From 2D to Boris Vallejo Paintings and other artists this will do almost everything with a bit of prompt crafting!

It also works with LORAs & EMBEDS without difficulty and generates some very unique and interesting results. (Can't guarantee every LORA will work with it. but many I've tried seem to without issue.)

It's created to be a catch all for this particular medium and I've tested this model for a month exclusively and have been very impressed with its results.

No embedded VAE but this particular vae is required to get the results & examples shown: (The Vae has also been added to the file list)

YOU NEED TO USE THIS VAE!

☞vae-ft-mse-840000-ema-pruned ☜for v2.5

☞kl-f8-anime2.ckpt ☜ for v2.0


2.5 is more utilitarian in that it specializes now in both oil painting & flat shaded 2D & everything in between!



🔥PRO TIPS & TRICKS:🔥
====================================================

Works best from 4-15 CFG and up 20-150 STEPS! YMMV!

There is a PRUNED (COMING SOON!) & FULL VERSION

=============>🌟WORKS BEST WITH CLIP SKIP 1! 🌟<===========
(seriously use this or you will not get results like the image examples at all.)

I recommend & use the following negatives commonly:

EasyNegative (The all around catch all. Always good in a bind)
badhandv4_anime (absolutely your go to neg for anime / drawn hands)
bad-hands-5 (good for the more realistic images but many times will not work amazingly for just anime. I usually use this and badhands4_anime together)
bad-artist \ bad-artist-anime (I've had good & bad luck with this but sometimes it's great.)
bad_prompt \ bad_prompt_version2 (works very good sometimes)

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

This is complex mix that I'll share information about how it was made in the near future.
I'm going to take a break for a while as I've been working on this fairly solid for over a month straight! I hope you all enjoy it and show me what you make with this magical model <3


🌟Below (and also other images in the gallery) are examples of the styles you can achieve from the model just by changing the weights on a few simple prompts like: (Comic Style), (Realistic), (Line Art), (Illustration), (Curisscaro), etc. (These image examples w/ prompts are in the gallery for your use & reference.)

MANY examples will be going up in the next few days. So keep your eyes peeled!

=========================================================
A walkthrough on how to setup and use the basics in A1111 to get the results you see:
===========================================================

Blurred eyes, or sometimes warped faces are not that uncommon in SD and sometimes require an inpaint to pull off. Most of my examples are not inpainted. But a few are.

First off when you run txt2img ALWAYS use HIRES FIX! <===

Upscaled: latent (nearest or exact) upscale, or 4x-UltraSharp.
Denoising: 0.35-0.7 denoise (Complete variable - depends on the type and complexity of the image. YMMV! But 0.4-0.55 is a good window.)

Upscale by 1.5-2 but no more than that.

This will get you a pretty good base image to work with.

Afterwards.. You can always send your txt2img picture to inpaint on the bottom right under the result area > select the face or eyes > lower the denoise strength to 0.5-0.57. > make sure your seed is set to random.

masked content = fill / inpaint area = whole picture ====== for faces

masked content = fill / inpaint area = only masked ====== for eyes


With "whole picture" selected for larger inpaint target aka faces (or hands) you may need the denoise to be lower to pick up more of the original image. 0.53-0.56

Now with "only masked" and a smaller target like eyes you get a little more freedom so your denoise can be 0.54-0.57

Run a batch of 4-5 (or more) for the faces.. go through pick the best one.. if you're lucky you got the eyes and face in one go... if not you may need to choose the best face.. send back to inpaint again.. select just the eyes.. and rerun a new batch (changing it to only masked and the settings described above for eyes.)

Finally when you get it all together send to extras:

Resize to 1.5-2

Upscale 1= Nearest
Upscale 2= R-ESERGAN 2x or 4x

But google or get yourself these:
Remacri Upscaler
Lollypop Upscaler
4x_Fatality_Comix_260000_G
4x_NMKD_Superscale-SP_178000_G

(Install these into the proper directory: A1111-Web-UI-Autoinstaller\stable-diffusion-webui\models\ESRGAN and reload a1111.)

Upscaler 2 is what will get you the results you want. You can also drag the upscaler 2 visibility to control the amount of effect each puts on the finished result. (I recommend turning it up 1 but YMMV.)

Ramacri / Lollypop / & Fatality Comix are neck and neck for the best output!
They are all different. One is sharpest, One is more subdued, and one is like a sharp HD.
Each are good for any type of comic / illustration / cell shading / anime.

NMKD_Super is more for semi-realistic to real images and good for scenery. Tends to run sharp so you may have to turn it down a bit.


Each of these is unique and so is their outputs if you want to compare simply change the pull down and run off the various types and contrast / compare all your the finished results.


For the more advanced upscale: Get yourself Ultimate SD Upscale
(click the link for a basic tutorial on how to get it and how to use it.)

Good Luck & Have Fun!

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

this is a Blend of EveryDayArt [EDA] (a checkpoint I am still working on), CounterfeitV3, MenaiHentaiV3 and Paizuri FocusV1 (can't remember the exact percentages - but if I find them I will put them) [so it's half original and half merge, not that that matters, just thought it was cool]

this is an illustration/anime model, can do nsfw and sfw, and gives good paizuri images if that's your thing.

for the paizuri stuff visit my model and/or my LoRas - Paizuri Focus - Paizuri-FocusV1 | Stable Diffusion Checkpoint | Civitai

for everything else, well if your downloading this you know the drill, get creating.

found bugs: pov likes to result in paizuri unless fought or your more specific (Ie: put pov in both negative and positive and/or specify more of the scene like doggystyle, from_behind, pov, etc), and even then I found that genning in batches usually only the first or first two images would be paizuri, but could have just been lucky

note - some of the images were used with orangemix Vae, so that's why some are less saturated, extra note - think this is wrong, was going to use those images but then didn't, also vergil model is burnt, not a good showcase of the quality. extra extra not - yes first note was right, some did use orangemix

authors note - I am terrible with names, would like any that you guys come up with, preferable blood or red related