[Example images went through 'SDupscale' process.]

[So even if you copy exif, Same image may not be generaited.]

This model generait cute face girl

Also generait sexy body girl

▼Please check settings and prompts▼

Merge model

Anime Full

Treebark

Counterfeit_V2

CetusMix_Version3.5

CetusMix_Codaedition

CamelliaMix_2.5D_V2

HongBsWs style LoRA

Tabi Art Style LoRA

Project Sekai Style LoRA
Neco style LoRA

Recommended setting

Sampling method : DPM++ SDE Karras, (or DPM++ 2M Karras)

Clip skip : 2 

Hires steps : 13

Hires.fix upscaler : R-ESRGAN 4x+Anime6B (or R-ESRGAN Genaeral WDN 4xV3) 

CFG Scale : 7~10

VAE : kl-f8-anime2 (or Orangemix.pt)

Recommended prompt

Prompt : (masterpiece:1.2, best quality)

Negative : (worst quality, low quality:1.4), negative_hand-neg, verybadimagenegative

Textual Inversion : negative_hand-neg, verybadimagenegative

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.

NOTE: 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 :)

☕You can buy me coffee

If you support me, I will make many AI male models and try to improve them more beautifully.

***

BlueBoys Ver.3 can create more detailed and manly male characters. It's more of an illustration than an Anime2D.

Ver.3 makes high-quality illustration, but if you want to make a simple style illustration or various characters, it's still good to choose Ver.2.

Recommended Settings

• Sampling method : Eular a / DPM++ SDE Karras

• Clip skip : 1~2

• Hires.fix upscaler : R-ESRGAN 4x+Anime6B

• CFG Scale : 7~11

• VAE : vae-ft-mse-840000-ema-pruned / kl-f8-anime2

This model has been tested to be effective in producing AI Cosplay Characters with Control Net. Particularly female human cosplayers can be more easily generated. Ionising strength optimal between 0.22 to 0.44. 0.22 for small details to be captured then inpaint the rest of the image with 0.33 for closer to original art lines or 0.44 for more realism.

Works well with realistic prompts.

Samples I shared are mostly Image to Image results.

Use Control Net models to get closer results.

Please do not use this model for illegal means.

Thanks and hope you find it useful!

many many checkpoints merged , including 2 of my own custom dreambooth creations..

A 2.5D model I merged based on majicMIX_lux. It has a similar approach as majicMIX_fantasy but is not directly related, so I named it "reverie". You may figure it out through my examples.

试试这两个插件 I used mcmonkeyprojects/sd-dynamic-thresholding: Dynamic Thresholding (CFG Scale Fix) for SD Auto WebUI (github.com)

AND Bing-su/adetailer: Auto detecting, masking and inpainting with detection model. (github.com)

works extremely well with zhongfenghua style which is definitely undervalued

Version 3 is here, hope you have fun with it.

The model is better with a VAE. (vae-ft-mse-840000-ema-pruned is fine or you can use your own.)

* For some reason (for V2) the pruned version isn't showing as the first download and the hashes have got screwy. Please use the dropdown menu to choose the smaller version if desired *

This model is a custom blend of various models, (see below), giving many options for generating a wide variety of images, including NSFW. Based on Stable Diffusion 1.5.

Primarily focused on capturing skin texture and the feel of photography, and Polaroids in general, but experiment with other things. (This model is not perfect).

Though there are is no specific KEYWORD to 'turn it on', if you use words in the prompt such as 'polaroid' , 'analog' , 'film grain', 'perfect eyes' and so forth at the start of the prompt, you should get the look you're after.

The example pictures do not have any face/eye correction or upscaling. 768 x 1024px is quite stable at producing images without 'mutation' if the prompt and neg prompt are good. Version 2 will hopefully address the eye issues and understanding the age input better.

The model is better with a VAE. (vae-ft-mse-840000-ema-pruned is fine or you can use your own.)

Prompt example:

(polaroid), portrait, ((extreme realistic skin detail)), (20yo) (SUBJECT NAME or PERSON TYPE), ((holding a heart balloon)), lingerie, sassy, short black hair, sharp focus, detailed skin, detailed face, skin pores, skin imperfections, film grain, dynamic lighting,

Negative:

child, blur, old face, extra arms, big chin, (long neck!), ((more than one face)), imperfect eyes, glowing eyes, (head crop), pout, (abnormal body), double head, oversaturated, (head out of frame), too close, bad hands, distant, ugly, male, fake breasts, tiling, large nipples, extra heads, deformed, cgi, anime, painting, boring pose, plastic skin, extra limbs, big forehead, abnormal, long body, long limbs, overexposed, text, watermark, hairy body

A non-exhaustive lost of models merged (because I didn't take notes): deliberate, clarity, dreamlike-photoreal, realistic vision, protogen (+ additional ones in version notes).

Made for fun, not for money or notoriety. Those who train the models from scratch are the ones deserving of applause.

Have fun! Can't wait to see what you come up with...

Posts of images welcome.

Please feel free to leave a review about the model.

If you would like to see more consider supporting me you can do it here:

https://www.buymeacoffee.com/goldenwraft

https://www.patreon.com/GoldenWRaft

Used as base models:

Clarity:V2: https://civitai.com/models/5062/clarity

ConsistentFactor:V4.0: https://civitai.com/models/9114/consistent-factor

Colorful:V2.6: https://civitai.com/models/7279/colorful

Do you want to use artists?

https://stablediffusion.fr/artists

• Adding [by Artist, etc] at the end of the prompt can give you the desired result

Nicknamed: C3

Credit to justmaier

Version info can be found under the review section.

I hope you enjoy.

!NB Please do not use this model for NSFW purposes! I highly recommend placing sfw in your positive prompts.

Also available at 🤗: https://huggingface.co/Miyuutsu/kittenchow

This model was originally a dreambooth training. It has since been through more merges with more models than I care to remember or count and has been fine tuned with images from various artists meeting the style. I have marked this as "Trained" because, despite it containing merges, it did in fact have a good amount of training too.

It's made to be incredibly cute and adorable while being capable of taking on basically any Lora with some form of efficiency. I would recommend using lower weights for your Lora with this model than you would any other model. This model also does very good scenery!

It is unlikely you can perfectly replicate my previews as I am using ToMeSD and various custom configurations.

A style + merge model

Adjustments have been made to VAE, but you can try other VAE as well

characteristics

↓

The model performs better on the female body

About hand deformities

↓

Maybe it can be corrected by controlnet, or maybe I'll get a new one later

LoRA's adaptability

↓

generally, we tried a few Genshin characters, but they were not as amazing as I thought

Cute Anime semi-realistic creepy art

Took me a while with my blending machine to create this checkpoint. capable of producing a wide range of artistic creations. From adorable anime illustrations to chilling horror scenes.

Positive prompts:

• (masterpiece:1.2), (high resolution:1.2), (best quality:1.1), (HDR), (cinematic lighting), (sharp focus), I'm using this as my base style.

• (creepy anatomy), (horror illustration) for creepy stuff.

Neg prompts:

• (worst quality, bad quality:1.3), (watermark) (label) (signature) (text), I'm using this as my base style as well.

I'm experimenting with different Prompt Embeddings and Lora, so far looks good!

I merged my favorite model.
It's a model to make a romantic couple. It is possible to express a very delicate character.

Recommended Sampler
DPM++ 2M Karras or DPM++ SDE Karras

Recommended Upscaler
Latent (Denoising 0.5 ↑)

Recommended Prompt
(masterpiece, best quality) or (absurdres, highres, ultra detailed)

Merge Material
Glow 2.5D
Victorian Mix
Detail Tweaker LoRA
GlitterMix
Counterfeit-V3.0

There should be many beautiful words why this mix is better than others, but there won't be any, the best it is only for me. There are a dozen LORA models assembled, anime stylization (I'll be honest with the hentai), one realistic model - heavily influenced this mix and at the end I added an emphasis on landscapes. The landscapes, even with minimal querying, turn out surprisingly well. And yes, don't be surprised if it suddenly gives you monsters.

I recommend using it in conjunction with: vae-ft-mse-840000-ema-pruned

https://huggingface.co/stabilityai/sd-vae-ft-mse-original/blob/main/vae-ft-mse-840000-ema-pruned.ckpt

AOM3 + 三餡馄饨Mix + 啥玩意完犊子

自用merge

Since I've learned about how to merge models and inspired by the results from NostalRealMix, decided to try to make my own just for a little fun.

So I've decided to merge BRA(Beautiful Realistic Asians) model with BB-MIX-EVE. The results are not bad.

Parameters I use:

• DPM++SDE Karras

• Sampling steps: 20

• R-ESRGAN 4x+ Anime6B

• Denoising Strengh: 0.45

• Hire steps: 10

• Upscale: 2

• Resolution: 512x768

• CFG Scale: 7

• Clip Skip: 1 or 2

A woman transforms into A giant Busty Anthro Dinosaur with huge giant Breasts, ass, hips and Thighs.

The continuation of a weird model!

Having used alfahentai for lot random art, it was only being held back by it's own confusion at times. This model is built upon the idea to get rid of that confusion and make a confident model. All the images i posted are only using Nerfgun3's Negative Hand embedding. The rest is pure txt2img prompt. No inpainting or other manipulation was done.

Big shoutout to Cryonicus, I want to believe his model was the missing part my model needed. Make sure to give his model a try too!

What this model is and isn't

Unfortunately, getting the art to improve to an incredible standard, it cost the model it's ability to make proper genitalia. Breasts still work though. Again, the model has been merged as much so that it retains it's general knowledge. It does NOT respond well booru specific tags/prompts. It might have recollections of some things you might expect in a lot of booru tagged models. For example Wariza doesn't function. Masterpiece has essentially been removed as tag and is generally not needed at all. For all the images I create, unless I stole someone's prompt, it will not have masterpiece/best quality ever. Honestly, i've tried comparisons and generally doesn't seem to add anything.

This model shines in it's ability to grasp concepts, context and general cohesion. I've tried various prompting styles on this model and they all seem to work. It works best with natural language and descriptions.

Model Usage

The model works best on a low CFG, i stick at 5. At 6 you already start getting the first symptoms of burn, at 7 it's really bad, and 8 and above simply seem unusable. The model has a lot of detail it can resolve, so it likes high stepcount, like it's predecessor. With UniPC and cfg 3 or 4 you test prompts with steps as low as 7. Take note, that sometimes your prompt will be interpreted differently if you change the aspect ratio. The unpruned image has a bit better image quality in my experience.

CFG: 5

Steps: 20-50

Clip skip: 1

Negative prompt used

negative_hand, (low quality:1.3), (worst quality:1.3), (lowres:1.3), (error:1.3), (fault:1.3), (undefined:1.3), (crude:1.3), jagged, spiky, ugly hands, extra arm, extra hand, split arm, missing finger, extra finger, three fingers, four fingers, six fingers, (disfigured:1.3), (unclear:1.3), (indistinct:1.3), merged fingers, (bad anatomy,:1.3) misplaced hand, misplaced foot,, cleavage, sex, nsfw, ecchi, slutty, leotard, (text:1.3), (signature:1.3), (title:1.3)

类似于上色漫画

Similar to coloring comics

好吧给个范例，我的目的，大概可以这么做

Well, to give you an example of what I'm trying to do, you could probably do it this way

先写好剧本定好分镜吧，固定一下设定tag

Let's write the script and set the storyboard, fix the tag

下一步就是出图了 The next step is to map it out

然后合成，嵌字 And then compositing, embedding

最后就是例图的效果 Finally, the effect of the example diagram

ღ( ´･ᴗ･` )比心

Steps: 20, Sampler: Euler a, CFG scale: 5, Size: 512x768, Model hash: 669eeb48a5, Model: 23511-1546-幻色石改fp16-fp16-no-ema, Denoising strength: 0.5, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+

This Checkpoint can use for Anything you can imagine

(work for 2D/2.5D/3D/Anime/Hentai/Artwork with lora)

Prompt

• (General using prompt)

Negative

• (EasyNegative, badhandv4, ng_deepnegative_v1_75t,)

Default Setting.

DPM++ SDE KARRAS

Sampling: Steps 20-40

Upscaler: 4x-UltraSharp

Hires Steps: 10-20

Denoising Strength: 0.4-0.6

CFG Scale: 7

Mimic CFG Scale: 7.0

Clip Skip: 2

Recommended VAE: ClearVAE or vae-ft-mse-840000-ema-pruned

** check my another checkpoint Asian Mix **

high quality anime style model.

more info. https://huggingface.co/AerinK/NotSoXJB-Mix-1
EasyNegative(Negative Embedding) https://huggingface.co/datasets/gsdf/EasyNegative

badhandv4(Negative embedding)

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

baked with vae. or you can use your own vae.

Merge various checkpoints

VivCharMix - Effect of transforming ACG characters into real-life people is outstanding

Recommend Negative Words: (low quality, worst quality, lowres, grayscale, normal quality, username, outline, poorly drawn, copyright name, character doll, scribble , silent comic:1.4)

(The English translation can be found below the text)

這個模組將ACG角色轉換成現實人物的效果非常出色

如同從圖畫中活生生過來，

避免像Coser(真人對應不足、外表太過真實)、人偶(眼過大、小鼻子)、手辦(表情僵硬、塑膠感)，使生成的真人角色更貼近原作

這個是經過微調的比例，可以有更好的寬容性和結果
EasyNegative 越高臉型越符合，但畫面會越銳利

verybadimagenegative_v1.3 越高衣物細節越正確，但像油畫

sd1-failed-oekaki 越高背景越好，但越像小孩而且嬌小

可以視需求微調

因為所有微調都是建立在這組負面詞上

若不使用這組負面詞，無法達到我煉製和分享的模組預期

不需要VAE

優點：
(1) 臉部好看的生成模型，適合生成ACG人物真人化，保留氣質和神韻和衣物細節

(2) 對ACG的Loras有高適應性和寬容度(建議權重0.3~0.8，用0.05當作一個區間調整)

(3) 鮮明飽滿的顏色和動態的表情

(4) 可以相對正確的生成NSFW內容，以及可正確表現體液

(5) 對畫風style類Loras敏感，大概0.3就足以改變輸出結果，0.6會大幅影響畫面，增加可玩性。 對姿勢類、服裝類Loras也有相當好支援性

(6) 可以產生ACG中出現的裝飾元素，例如spoken heart、emphasis lines、speech bubble...

--> 使用新的負面詞後，甚至可出現紅眼、綠眼

(7) 適合當作ACG轉真人圖生圖的模型

負面詞：EasyNegative, verybadimagenegative_v1.3, sd1-failed-oekaki、修改幅度0.45

經過兩次img to img

缺點：

(1) 場景單一，特別容易出現日常或素色背景，建議用Loras或Prompts調整

解決，加上Oekaki negative TI負面詞後，會擁有更出色的品質和印象深刻的背景

(2) 沒有對ACG男性角色微調，可能導致娘化

好看的阿福

(3) masterpiece關鍵字被錯誤綁定，容易生成nsfw內容

(4) 使用新關鍵字後，身材過瘦、過於平坦，我覺得如果更凹凸有致
並且有點肌肉會更好看。 或著可以用prompt或Loras調整

目前找不到更好的相同的model可以merge，在我能學會mwb之前這個就是完成版了

特別感謝：
KanPiroMix

3moon anime_cosplay

3moon REAL_Photo

DDosmix

Lofi V2

與其他真人化效果

原圖 Original image

加上新的負面詞Oekaki negative TI重新比較

原圖2 Original image 2

(Translated by Chatgpt)

VivCharMix

This module's main purpose is to transform ACG characters into real-life people.

It brings characters to life from drawings while avoiding issues like cosplayers (insufficient real-life representation, appearance too REALISTIC), dolls (oversized eyes, small noses), and figurines (stiff expressions, plastic feel).

This is a fine-tuned proportion that offers better tolerance and results.

The higher the EasyNegative, the more accurate the face shape, but the image will be sharper.

The higher the verybadimagenegative_v1.3, the more accurate the clothing details, but the output will resemble an oil painting.

The higher the sd1-failed-oekaki, the better the background, but the result will look more childlike and petite.

You can fine-tune according to your needs.

All fine-tuning is based on these negative Prompts. Without using them, it's impossible to achieve the expected results of the module I created and shared.

No need for VAE

Advantages:

(1) Generates good-looking faces, suitable for turning ACG characters into real-life people, preserving their looks, charm, and clothing details

(2) Highly adaptive and tolerant to ACG Loras (recommended weight: 0.3~0.8, adjust in 0.05 intervals)

(3) Vivid, full colors and dynamic expressions

(4) NSFW Okay!! Visable Body Fluid!!

(5) Sensitive to style-related Loras, roughly 0.3 is enough to change output results, 0.6 will significantly affect the image, increasing playability. Also offers good support for pose and clothing Loras

(6) Can generate ACG decorative elements, such as spoken hearts, emphasis lines, speech bubble, etc.

--> After using the new negative terms, even red eyes and green eyes can appear.

(7) Suitable as a model for creating real-life images from ACG characters

Recommended negative Prompts: (EasyNegative:0.9), (verybadimagenegative_v1.3:0.9), Amplitude: 0.45

-

After two img-to-img adjustments,

-

Disadvantages:

(1) Limited scenes, especially prone to everyday or simple backgrounds, recommend using Loras or Prompts to improve

Solved. By adding "Oekaki negative TI" negative terms, the quality and backgrounds will be even more outstanding.

(2) No fine-tuning for ACG male characters, maybe looks girly

Good-looking Astolfo Face

-

(3) The "masterpiece" keyword is mistakenly bound, easily generating NSFW content

(4) After using the new keywords, the body is too thin and overly flat. I think it would look better if it were more curvy and had some muscle. Alternatively, you can adjust it using prompts or Loras

Currently, no better, similarly purposed models can be merged. This is the completed version until I learn MWB.

Special thanks to:

KanPiroMix

3moon anime_cosplay

3moon REAL_Photo

DDosmix

Lofi V2

Compared to other real-life transformation effects

A merged model to generate images of 2.5D cute girls

Please use simple prompts and negative prompts.

Модель включает в себя лучшее, на мой взгляд, сочетание реалистичных стилей. Она соединила в себе более 15 различных моделей в четко выверенных пропорциях, что позволяет получать гипер реалистичные портреты невероятного качества. Модель натренирована на многих знаменитостях, рисует их лучше, чем любая Лора или Ликорис. Также содержит реалистичных nsfw контент.

The model includes the best, in my opinion, combination of realistic styles. It combines more than 15 different models in well-adjusted proportions, which allows you to get hyper realistic portraits of incredible quality. The model is trained on many celebrities, draws them better than any LoRaor Lycoris. Also contains realistic nsfw content.

Please select the version you like, not the latest.

My miscellaneous models.

If you would like to see the recipe or more detail, feel free to ask :>

recipe missing:

(03191422,03192033,03191454)

SweetBoys 2D is a model that can make slender and cute boy characters.
Fall into the very fantastic and delicate world of beautiful boys!
This model tends to ignore Prompt in Clip skip 1 and draws an abstract illustrations. Clip skip 2 or higher is recommended.

***

☕You can buy me coffee

If you support me, I will make many AI male models and try to improve them more beautifully.

This model can create more delicate and fantastic female and male characters.

It is recommended to use the Detection Detailer extension to make your face beautiful.

Ultra Realistic Porn

Lots of Pornstar include

My Fans Lora : fanslora.com/miaou
Trained with +5k HD Pictures (20% Merged PoV Skin Texture)

Here is the style of prompt : https://miniurl.be/r-4k0j ( you can mix prompt )

This is a beta version, an update is coming soon with corrections bad body result.

By the way,it is still possible to have some nice result, have fun and don't hesitate to share your images ! 

Damagde version

This model was trained on artwork that had glowing affects primarily done in pencil of portraits.

Created in Dreambooth this model provides you the ability to create extremely detailed 2D Illustration pencil Styles.

I have use many prompts from Lexica.art as well as from arthub.ai that have given me some really phenomenal results.

Making these models can be expensive and very time consuming. If you would consider helping out, it would be much appreciated as any little bit goes a long way

https://www.buymeacoffee.com/Theprotoart

I also show you how to create your own custom models with your own art on my YouTube channel.

My Very Own Model

Just a small fusion of Somethingv2 and RuviMixV1, still a working progress but I will someday make it work

Making Children's Books, Comics, Graphic Novels is rewarding, but very heavy on the artwork. This model will help you tell any of your stories visually. Let your imagination go wild. If you need custom characters that are consistent throughout a story, you can hire me to make one for you. Contact me in one of the links under my profile.

PROMPT RESPONSIBLY

Two models included:
V1 Toon-Anime and V1 Semi-Real. You won't be able to make photorealistic images with this model but that's not the point. Create colorful, fun, and imaginative images that excite a child's wonder.

Do you have requests? I've been putting in many more hours lately with this. That's my problem, not yours. But if you'd like to tip me, buy me a beer. Beer encourages me to ignore work and make AI models instead. Tip and make a request. I'll give it a shot if I can. Here at Ko-Fi

A new blend I'm testing. I use it often for lora testing so I thought I would upload it.

Here is my base model. FitCorderMix. My realistic mix is fCBlendMix. My anime mix is fCAnimeMix.

Check out my other Lora's.

Body Type: Double Thick Thighs, Perky Breasts, Round Breasts, Max Cleavage, Amazon Goddess,

Fashion: Crop Tops, Choker Minidress, Scrunchbutt leggings, Cybergoth, Nerdy

Environment: Undersea Depths

A new blend I'm testing. I use it often for lora testing so I thought I would upload it.

Here is my base model. FitCorderMix. My realistic mix is fCBlendMix. My anime mix is fCAnimeMix.

Check out my other Lora's.

Body Type: Double Thick Thighs, Perky Breasts, Round Breasts, Max Cleavage, Amazon Goddess,

Fashion: Crop Tops, Choker Minidress, Scrunchbutt leggings, Cybergoth, Nerdy

Environment: Undersea Depths

Just another merge of some popular models, that I think turned out good.

It's somewhere between 2.5d and realistic. The merge is a bit biased towards nudity, blonde hair and Asian looks, but is still pretty versatile. (see example images)

There is no VAE baked in, so I suggest using either vae-ft-mse-840000-ema-pruned.vae.ckpt or kl-f8-anime2.ckpt with the merge.

The merge works well with either clip skip 1 or 2.

Updated models in v1.1:

uberRealisticPornMerge_urpmv13

henmixReal_v30

cyberrealistic_v21

revAnimated_v122

deliberate_v2

This is merge of chilloutmix and henmix real

0.6(undercucho) + 0.4(uberRealisticPornMerge_urpmv12)

Prompt Close portrait of elegant [undercucho] person in tailored suit- futurist style, intricate baroque detial, elegant, glowing lights, highly detailed, digital painting, artstation, concept art, smooth, sharp focus, illustration, art by wlop, mars ravelo and greg rutkowski (high detailed skin:1.2), 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3 <lora:breastinclassbetter_v141:0.5> <lora:inniesbettervaginas_v11:1.0> mix4

Negative prompt

anvas frame, cartoon, 3d, ((disfigured)), ((bad art)), ((deformed)),((extra limbs)),((close up)),((b&w)), wierd colors, blurry, (((duplicate))), ((morbid)), ((mutilated)), [out of frame], extra fingers, mutated hands, ((poorly drawn hands)), ((poorly drawn face)), (((mutation))), (((deformed))), ((ugly)), blurry, ((bad anatomy)), (((bad proportions))), ((extra limbs)), cloned face, (((disfigured))), out of frame, ugly, extra limbs, (bad anatomy), gross proportions, (malformed limbs), ((missing arms)), ((missing legs)), (((extra arms))), (((extra legs))), mutated hands, (fused fingers), (too many fingers), (((long neck))), Photoshop, video game, ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, mutation, mutated, extra limbs, extra legs, extra arms, disfigured, deformed, cross-eye, body out of frame, blurry, bad art, bad anatomy, 3d render

Please Read Description

Questions/Feedback/Updates?

Visit my thread on the Unstable Diffusion Discord

Description

Built with 1980s - 1990s fantasy/sci-fi illustrations in mind but is capable of much more.

This model is a "prequel" to a much larger custom trained model I'm working on. The future model is trained will be trained on ~1k manually captioned fantasy/sci-fi illustrations from the 1980s-1990s. The future model will be named "Boomer Art Diffusion"

The base model for Electric Eden (this model) is an undertrained sample of Boomer Art Diffusion. Being undertrained, I merged a few models with the base to fit the model properly. A list of the models used in the merge can be found in the "About this version" tab.

To get the classic image style use :

• Artist names from the era

  • Boris Vallejo, Frank Frazetta, Julie Bell, Jeff Easley, Larry Elmore, Luis Royo, Olivia de Berardinis, Jean Henri Gaston Giraud (or) Jean Moebius Giraud

  • A full list of artists will be available when the full model is released.

• Heavy Metal Magazine Cover art <1977-1996>

• Illustration

• Dates are important. Try adding a year somewhere in the prompt.

  • Notably 1980 to 1996

For Modern 3D renders :

• Nothing special, prompt as usual for this type of generation.

For Anime Style :

• (Anime Style)

  • Feel free to adjust the weights as needed, example (Anime Style:1.2)

A version of this model that uses Booru tags can be made if requested enough.

Important Settings / Recommendations

VAE :

• vae-ft-mse-840000-ema-pruned.safetensors

Note : you can use the pytorch (.ckpt) version if needed

Install :

stable-diffusion-webui -> models -> VAE

In Webui :

Under Settings -> Stable Diffusion -> SD VAE select the downloaded VAE and uncheck Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them

Negative Embeddings :

• bad_prompt_version2

  • Weight in samples images : (bad_prompt_version2:0.8)

• bad-hands-5

• EasyNegative

How to install :

place .pt files in stable-diffusion-webui -> embeddings

Note : Negative embeddings are not required, but were used in the creation of the sample images. You can change the weight (overall affect on output) by using Attention/emphasis for example; (easynegative:0.5), (bad-hands-5:0.6)

Webui Settings :

• Clip Skip : 1

  • Settings -> Sampler parameters -> Clip skip

• ETA Noise Seed Delta : 31337

  • Settings -> Sampler parameters -> Eta noise seed delta

• Face Restoration : Disabled

  • See "Adetailer" for more info.

  • IF enabled : Settings -> Face restoration -> CodeFormer -> weight = 0.8

Parameters :

• Sampler :

  • I used a combination of Euler a, DPM++ SDE Karras for many of the sample images, but this is just personal preference.

• Steps : Depends on the sampler, typically 20 - 35 works well.

  • If DDIM use higher steps

• CFG :

  • Electric Eden v1 - 6.5 - 10

  • Electric Eden v1 Dark - 5 - 7

Hires Fix (Optional) :

• Upscaler :

  • R-ESRGAN 4x+

    • Denoising strength : 0.3 - 0.35

    • Upscale by : 1.5 - 2 (can go lower but shouldn't go higher than 2)

  • Lolipop, 4x-UltraSharp, Remacri (foolhardy)

    • Can be downloaded here

    • Install to stable-diffusion-webui -> models -> ESRGAN

    • Typically Denoise ~= 0.5 - 0.6 (can go higher or lower)

    • Upscale by : 1.5 - 2

Adetailer :

Extension for webui that uses ultralytics to 'fix' faces/hands/body after generation.

Download here - https://github.com/Bing-su/adetailer

Read the information in the Github repo linked above for more info and installation guide.

As for settings - I will need more experience with this extension before I can recommendations.

Check out my other models (also the models used in the merge)

• Cine Diffusion - https://civitai.com/models/50000/cine-diffusion

• ProjectAIO - https://civitai.com/models/18428/project-aio

• Eris - https://civitai.com/models/21952/eris

• WonderMix - https://civitai.com/models/15666/wondermix

• Refined - https://civitai.com/models/8392/refined

• Experience - https://civitai.com/models/5952/experience

• Elegance - https://civitai.com/models/5564/elegance

• Clarity - https://civitai.com/models/5062/clarity

• VisionGen - Realism Reborn -https://civitai.com/models/4834/visiongen-realism

LoRA

• Pant Pull Down - https://civitai.com/models/11126/pant-pull-down-lora

The end

Stablegram

my fanslora : fanslora.com/miaou

US/European instagram woman style

Warning: Add "nude, naked" to the negative prompt if want to avoid NSFW.

V1-V2.2 : (Chilloutmix base) Trained with +70 000 pictures on EveryDream2

V2.3 : BasilMix base Trained with +70 000 Pictures merged with PoV Skin Texture - r34 Lucid Black + PoV Skin Texture - Dreamlike r34

• Prompting: Simple prompts are better. Large amounts of quality tags and negative prompts can have negative effects.
  

Currently working on a new version with better accuracy and consistency. By the way, have fun and don't hesitate to share your images ! 

Artistic version of rMadArt. VAE included, no need.

I'm trying to incorporate better color expression and characters that fit my aesthetic on the basis of "counterfeit" universality. Have fun playing!

recipe:

• 0.4 counterfeit30

• 0.6 (mix5)

  • 0.4 blackcat

  • 0.6 (mix4)

    • 0.4 navelorange21

    • 0.6 (mix3)

      • 0.6 (mix2)

        • 0.8 pastelmix

        • 0.2 colorbox

      • 0.4 (mix1)

        • 0.8 piledream

        • 0.2 cuteyukimix30

This model is an Unstable Diffusion community project and wouldn't have been possible to train without their generous help and resources. A deep thank you to them as well as everyone who's contributed data, their time, and their experience!

Recommended (but subject to change)

Positive: masterpiece, best quality, high quality, realistic

Negative: worst quality, low quality, anime, digital illustration, 3d rendering, comic panel, scanlation, multiple views, artist name, signature, error, text, cropped

There were no watermarks in the training data, that tag won't do much. As well, you don't need to beg it for high detail, photo-realism, etc. It'll do that right out of the box, so save yourself some token space and prompt for what you want to see.

This was trained on top of a merged base model consisting of: Grey Model, seek.art MEGA, Bara Diffusion, A Certain Model, RPG v4, and the Artstation model. These models were mixed using Bayesian Merger.

~120k images were captioned using a combination of BLIP-2 sentence captioning, booru tags, quality modifiers, and the sfw/risque/nsfw tags (depending on the content of the image). Booru tagging was done first, and that result was used to help condition the output of the BLIP-2 captioning. This helped with coherency and gave it a much needed lewd tilt.

It was trained for 30 epochs using the EveryDream 2 fine-tuner, a base resolution of 768 (so lower resolutions aren't recommended when prompting), a batch size of 10, clip skip 1, zero offset noise set at 0.2, text encoder training turned on for three epochs, and a learning rate of 7.7e-7.

If you use this model as part of a mix, please mention and link back to it :)

We'll be entering a slow-burn period as data is manually tagged by volunteers for further fine-tuning. If you're interested in contributing (doing grunt work for me), you can join here. Please note, it's an offshoot of the UD community which is focused on this project so their rules apply as well.

This is an attempt to create a more "western" character look (or at least a non-standard anime one), as I've grown a little tired of seeing the same anime faces. Still very much a slightly NSFW-prone female focused character model though.

Does not require complex prompting - your mileage may vary prompting like you would for a more traditional anime model. Try building prompts from the ground up on this one. I've found its fairly compatible with most (non-style) LoRAs.

Somewhat unstable, so feedback would be greatly appreciated.

• All images generated using HiRes Fix with a denoising strength of 0.35 and a scale factor of 2.

• Would recommend EasyNegative and (worst quality:1.4), (low quality:1.4) in the negative prompt.

• VAE also recommended

This model was trained on the dataset with kawakaze images and then merged with other my models.

The model itself can produce other characters besides kawakaze, but sometimes there is a strong bias in the character traits, so I decided to name the model that way.

All images in the examples used lora with multiplier 0.4, this lora is in my profile.

And also, all the examples are made with multidiffusion, so the original seeds are lost.

Model is a result of mix'n'matching different trained or merged checkpoint with little experimentation with discarding/removing some weights.

Main goal was to create model that primary focus is men and content creation related to homosexuality, male relationships and NSFW gay content.

It's experimentation done just to learn before making a proper trained checkpoint.

Version 2 is based on my checkpoint merge and has been trained on additional data set containing 96 images.

Still working on fully trained checkpoint.

All generated samples has been made using Vladmandic or InvokeAI

If any picture has been upscaled I either used build-in upscaler in InvokeAI or UltraSharp x4

Non-standard resolutions i.e anything other than 512x512 or 768x768 has been made with Tiled Diffusion/Vae built-in Vladmandic's custom fork.

Anime tags and normal sentences work well, Unet & Text encoder has been trained to recognise both, although for best result you'll still probably guide model with increasing and decreasing weights i.e (word, tag or sentence)+ or - in InvokeAI or (word, tag or sentence:1.2)

Leave a review, and comment on what you wish to see to be improved or included in future releases, but please bear in mind this is a side project and a hobby so I can only do when I have any spare time.