I'm in the middle of practicing, so it might not work out well.

https://ko-fi.com/meg755291

english↓

This is lora by ishigami nozomi, a member of the virtual youtuber group nijisanji.

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

Trigger words:isgm,1 girl

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石神のぞみの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:isgm,1 girl

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。

So... They have done numerous porn out there, what's left to enjoy?

Well, some reasons maybe that they are retired, or getting old.. or you like younger version of them, or, maybe you just want to put them into a cosplay, a specific play that they never done in porn maybe? or you want to mix them with your other waifus and create some beautiful angel. Now thanks to SD, the possibilities is limited only your imagination and prompt, or maybe bad hands... or some mutated limbs, let us pray to the god of Ai that we are protected against such evil.

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

[EDIT : My model is out, check it: Tauron Hybrid Mix - TauHybrid_Anime_Final | Stable Diffusion Checkpoint | Civitai]

===
DISCLAIMER:

I'm doing this for fun only. my main focus is to generate a fictional character based on the beauty of all persons in this model. I do not support generating indecent imagery more than what they have been doing online, or selling generated image for commercial purpose.

If you love their work please support them, all models are 21+. Please DO NOT post explicit nudity, suggestive or pornographic content with the realistic image of the persons in civitai.

===

Open to suggestion or request

NOTE : I will keep updating more actresses in versions update

===

Usage Guide

Lora Weight = 0.1-0.4 (is best for combining, be mindful that your model may contain bias towards specific face such as korean or asian), 0.5-0.7 (is good to retain original face, while not too constraining your prompt), 0.8-1 ( is to generate accurate features of the original person ).

Use Dynamic CFG Threshold

• Mimic = 10 | Clamp percentile = 99.5

• Min. Mimic Scale = 0, Cosine Up

• CFG Scale Sch. = 15, Half Cosine Down

NOTE : I realized that with LyCoris Person, dynamic CFG actually lower success rate. without dynamic CFG result are more highly successful at the cost of a little less detail.

If subject face is small / far away use: DDetailer Script with bbox anime face with default settings except denoiser at 0.2-0.3. You can combine this LoRa with a lot of other Style / Attire / Pose / Acts LoRa, Also works well with character LyCoris.

Please leave a review if you have any critiques or if you find this LoRa good.

LyCORIS of Octavian "Kripparrian" Morosan

His youtube channel: https://www.youtube.com/@Kripparrian

His reddit: https://www.reddit.com/r/nl_Kripparrian/

His twitch: https://www.twitch.tv/nl_kripp

LyCORIS of Neil deGrasse Tyson

LyCORIS of Ranton

His channel: https://www.youtube.com/@RealRanton

His reddit: https://www.reddit.com/r/ranton/

Recommended weight: 0.8 - 1.0

Trigger: aoba kokona

LyCORIS of William Shatner

i buy Faceapp for training this dylora.

so... Leave your comment if you find it helpful.

How to use dylora?

1、This is a LyCORIS (LoCon/LoHA) model, and requires an https://github.com/KohakuBlueleaf/a1111-sd-webui-lycoris to work.

2、<lyco:Model:1:1:128>

you can change third number for any dim under 128 such as dim =1

<lyco:Model:1:1:1>

if you dont want to change background and pose ，just use low dim

Since I don't own a GPU and the only way to continue creating LoRAs is to rent GPUs (free Colab, prohibited Stable Diffusion, and remote UIs), I have chosen to do so with Paperspace. Please check out my Ko-Fi page if you are interested in supporting me.

This is a LoHa LoRA! You need the extension!
Why? LoHa have virtually no effect in output art style and they do a lot better with multi-concepts.

Trigger words for V4.14 - From 325 images to 910:

<lora:GirlyPenisv414:0.6>(futanari, penis, testicles)
You can now specify between "flaccid" OR "erection", "precum", "chastity, testicles" (don't prompt for "penis", "erection", "flaccid"), "bulge" (don't prompt for "penis", "testicles", "erection", "flaccid"), "penis out", "erection under clothes" (don't prompt for "erection","penis"), "foreskin", "tucked penis", "cock ring", "veins, veiny penis", "clothing aside" (can be reinforced with "leotard aside" or "panties aside" for example), "ejaculation, cum, projectile cum", "futanari masturbation", "femboy" (for more masculine face, probably doesn't work tho).

Its probably a good idea to complement the LoRA with more specific ones for things that its struggling with like "chastisty"... 70 out of 910 images had chastity in it but its still hard for it to draw that. I will make these specific ones if theres enough interest in it.
Most tags for camera angles and a few different sex positions should also work.

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

Sample images generated from CAN or Hassaku, check metadata!

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 910 512x512 images scraped from the internet and 14 epochs using animefull_final.

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

Adut Akech Bior is a South Sudanese-Australian model.

This is the LyCORIS version trained on the dataset I also used for the TI version (https://civitai.com/models/36954/adut-akech-model)

Please let me know which you prefer!

Let us embark on a journey to embrace some thot. Sometimes they are simply breath-takingly beautiful.

If you love my work and wish to buy me a coffe: ko-fi.com/tauron it helps me stays awake and get more content up

My model is out: Tauron Hybrid Mix - TauHybrid_Anime_Final | Stable Diffusion Checkpoint | Civitai

DISCLAIMER:
I'm doing this for fun and not for commercial purpose, I do not support selling imageries generated with all subjects inside this model. I do not support indecent generation images of all subjects used in this model. DO NOT post online explicit nudity, sexual or suggestive images using this model. All subjects are 21+.

===

Open to suggestion or request

NOTE : I will keep updating more actresses in versions update

===

Usage Guide

--

A. Lora Weight

• 0.1 ~ 0.4 -- is to combine multiple faces or features, or if you experience overbaking

• 0.5 ~ 0.7 -- is a balanced weight to get still accurate feature but with more flexiblity

• 0.8 ~ 1.0 -- is to create accurate representation of the features.

B. CFG Scale

• Lower CFG Scale means AI gets more creative and less overbaking problems.

• Higher CFG Scale means results get very accurate with training sets and more prone to overbaking problems.

• Dynamic CFG Threshold, using this method allows you to bump up quality and achieve photorealistic easily, but with lycoris i found out recently that this method actually lower success rate, so only use when you are composing with a Lora / model that works fine with this method.

C. Sampling Steps, High Res Fix & Upscaling

• I like to use sampling steps at 50, with high res step at 20, or 1/2 of sampling steps.

• I find that 2M Karras works better and faster than SDE Karras. with ERSGAN 4x+ method + ERSGAN 4x for Upscaling. This method actually gives quite good result without being too sharp or too plasticky

• Upscaling - GPFGAN @0.25, Codeformer @0.3 & Codeformer Weight @0.15 seems to yield superior results to me.

--

Combining with Attribute LoRa / LyCoris

You may need to do your own testing, but using character weight 0.2 below attribute lora weight yields superior results, if for some reason one features persists not showing or overbake, you may want to lower the weight of the overbaked feature Lora, and then use emphasis () to boost the specific attribute you want.

DDetailer Script

If subject face is small / far away use: DDetailer Script with bbox anime face with default settings except denoiser at 0.2-0.3. You can combine this LoRa with a lot of other Style / Attire / Pose / Acts LoRa, Also works well with character LyCoris.

How to install this script? I kinda forget how because its a bit complicated beacuse you need to install stuff from venv. but a quick google search will land you somewhere in tutorial page.

Please leave a review if you have any critiques or if you find this LoRa good.

Your dedicated onlyfans altar, summon your goddess here.

If you love my work, you can buy me a coffee to help me stay awake and cope with life to create more lora and stuffs : ko-fi.com/tauron

My Mix is out : Tauron Hybrid Mix - TauHybrid_Anime_Final | Stable Diffusion Checkpoint | Civitai

DISCLAIMER:

I'm doing this for fun and not for commercial motive, My main purpose is to compose a fictional character using AI with the beauty of all persons above, please support them if you love their works. i do not support any indecent images generated of all persons in this model, I do not support selling generated images of persons in this model. please refrain to what they have been posting online.

All persons in this models are 21+ and not a minor, please DO NOT post explicit nudity, suggestive content, or pornographic content online using this model.

===

Open to suggestion or request

NOTE : I will keep updating more actresses in versions update

===

Usage Guide

--

A. Lora Weight

• 0.1 ~ 0.4 -- is to combine multiple faces or features, or if you experience overbaking

• 0.5 ~ 0.7 -- is a balanced weight to get still accurate feature but with more flexiblity

• 0.8 ~ 1.0 -- is to create accurate representation of the features.

B. CFG Scale

• Lower CFG Scale means AI gets more creative and less overbaking problems.

• Higher CFG Scale means results get very accurate with training sets and more prone to overbaking problems.

• Dynamic CFG Threshold, using this method allows you to bump up quality and achieve photorealistic easily, but with lycoris i found out recently that this method actually lower success rate, so only use when you are composing with a Lora / model that works fine with this method.

C. Sampling Steps, High Res Fix & Upscaling

• I like to use sampling steps at 50, with high res step at 20, or 1/2 of sampling steps.

• I find that 2M Karras works better and faster than SDE Karras. with ERSGAN 4x+ method + ERSGAN 4x for Upscaling. This method actually gives quite good result without being too sharp or too plasticky

• Upscaling - GPFGAN @0.25, Codeformer @0.3 & Codeformer Weight @0.15 seems to yield superior results to me.

--

Combining with Attribute LoRa / LyCoris

You may need to do your own testing, but using character weight 0.2 below attribute lora weight yields superior results, if for some reason one features persists not showing or overbake, you may want to lower the weight of the overbaked feature Lora, and then use emphasis () to boost the specific attribute you want.

DDetailer Script

If subject face is small / far away use: DDetailer Script with bbox anime face with default settings except denoiser at 0.2-0.3. You can combine this LoRa with a lot of other Style / Attire / Pose / Acts LoRa, Also works well with character LyCoris.

How to install this script? I kinda forget how because its a bit complicated beacuse you need to install stuff from venv. but a quick google search will land you somewhere in tutorial page.

Please leave a review if you have any critiques or if you find this LoRa good.

This is a model of Armenian singer and songwriter Lilit Hovhannisyan. If you haven't heard of her, do check her out, she's pretty cool.

Please don't create anything disrespectful to her with my model, thanks!

Since I don't own a GPU and the only way to continue creating LoRAs is to rent GPUs (free Colab, prohibited Stable Diffusion, and remote UIs), I have chosen to do so with Paperspace. Please check out my Ko-Fi page if you are interested in supporting me.

This is a LoHa LoRA! You need the extension!
Why? LoHa have virtually no effect in output art style and they do a lot better with multi-concepts.

Trigger words for V1.20:

For original clothes: <lora:Ultronv120:0.8>(ultron)

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

Sample images generated from Hassaku.

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 35 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.

故 杉山新一氏の画風LoHA

人物や生物は学習させていないのでそれらの生成は推奨しません。

weight: 0.6-0.9

A request from net friend.

Role of Aigis.

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

TRIGGER:

1.BASE: original outfit, very long hair, low twintails;

2.Costume1: thighhighs, veil, bare shoulders, bridal gauntlets, garter straps, dress,covered navel,

3.Costume2:dress, thighhighs, black gloves, jewelry, earrings, navel cutout,ass visible through thighs, cleavage, clothing cutout, sleeveless blue dress, bare shoulders

非常非常简单的概念，可以自己配合写别的姿势。注意这是嘴吸胸部的概念，并不是固定构图。当然，最稳定还是普通的构图，但是你可以尝试结合一些可能的姿势。我在例图中仅给出一些简单的示范。

Very very simple concept, you can write other poses by yourself. Note that this is the concept of mouth sucking breasts, not a fixed composition. Of course, the most stable composition is still ordinary, but you can try to combine some possible poses.I only give some simple demonstrations in the illustration.

你要注意的是尽量不要写small breast

try not to write "small breast"

你可以尝试和其他姿势lora一起用，但由于姿势的契合度、lora质量不同，可能会得到不稳定的结果

You can try to use it with other pose lora, but you may get unstable results due to pose fit, lora quality

Trained on AOM3. This is a LoCON, so please install the appropriate extension to be able to use it. Then you can use it like you would a lora.

Consider tipping me through Ko-Fi!

https://ko-fi.com/muscleenjoyer

Recommended resolution: 512x768

lora weight: 0.8-1.0(I use Additional Networks)

Simpling method: DPM++ 2M Karras

Steps: 20

Model: "Chilloutmix_NiPrunedFp32Fix" or "Chilloutmix"

CFG: 7

No restore face

Statement

Production of NSFW works is prohibited

The model and its products shall not be used for commercial or profitable exchange

General instruction: use "red theme" to have hell, but hell can have others color too, up to you :D

Shit ton of leaking hair ornament :D

Sixth Lazy Style Lora.

Based on 582 image dataset of specific style.

Hell themed, obviously. 100% SFW, believe it or not, cuz again, MJ-generated.

Generally, this is not supposed to be used for anythin complex, because it was not tagged in a way to be, but you still can try and show me your results :3

Produces cute demon girls in combo with S12 <3 - https://civitai.com/models/22696/s12-v42-update-or-fine-tune-lora-or-extended-lora

GPTS Lazy Styles will have a lot of bleed, lower than average editability(but better than it was in GPTS1)

If you want it to be editable and still have some style effect, try weight from 0.5 to 0.75. That should make it editable and still be looking good :D
Maybe in this case you do want weight to be lower, as this particular one is usually abstract in backgrounds, and you will have hard time fighting leaks, because of dataset size.

Generally, this dataset processing used my custom tag filtering experiment, testing of which on GPTS1 led to increased editability, so, it might not be as hard to edit this one, but still, main purpose is just to get that bit of "wow" effect style.

Lighter version might be available later.

!! PLEASE READ !!

If the original artist requests a takedown, I will immediately remove this model off of CivitAI. Thank you. Please support their work here. Meanwhile, please do not monetize your generations.

Install the LoCon extension here. Or use that extension that's displayed in the blue box.

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

Weight: 1.0

Trigger: there's none

Try out sketch and flat color

These sample images do not contain (masterpiece, best quality:1.2) and the two negative embeddings I usually have.

LyCORIS of Henry Cavill

Dreambooth available at https://civitai.com/models/59622?modelVersionId=64070

LyCORIS of Elon Musk

LyCORIS of Charlie Cox

凸(☎︎ω☎︎)凸 凸(☎︎ω☎︎)凸

Kuso Model.

weight: 0.6-0.8

recommend prompt: chibi

キャプションは使っていないのでトリガーワードは不要ですがchibiと相性がいいようなので推奨します。

english↓

This is lora by kaburaki roko, a member of the virtual youtuber group nijisanji.

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

Trigger words:kbrg,1 girl

If you want her to wear a hat, add "hat" to the prompt to make it appear.

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鏑木ろこの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:kbrg,1 girl

彼女に帽子を被らせたい場合は、「hat」とプロンプトに加えると出現しやすいです。

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。

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!

LoHa of Yukino Bijin (umamusume). Please install extension on WebUI.

After installing extension, put LoHa file in your sd/models/lora folder and write LoRA notation in prompt to applying LoHa. <lora:your_loha_name:weight>

All of my works is free. If you'd like to support me, feel free to buy me a coffee. 👍

NOTICE

• After careful consideration, I decided not to produce LoRA of the characters below.

  • Otonashi Etsuko

  • Akasaka Misato

  • Hosoe Junko

• If anyone wants the characters above, I will consider making them.

• Except for the above, I will make the rest of the characters in umamusume unless there is a special notice.

Recommended options

• LoRA weight around 0.7

• Trigger words

  • Base

    • yukino bijin \(umamusume\)

  • Original Art :

    • ear covers, headband, white scarf, winter clothes, winter coat, black coat, long sleeves, fur trim, mittens, pleated skirt, blue skirt, miniskirt, zettai ryouiki, white thighhighs, white footwear, boots, ice skates

  • Racing competition costume :

    • ear covers, headband, white choker, white dress, long sleeves, fingerless glove, white glove, single glove, blue collar, necklace, buttons, brown belt, belt buckle, layered dress, frilled dress, white footwear, boots

  • Casual :

    • casual, ear covers, headband, collared shirt, pink sweater, cardigan, cable knit, long sleeves, blue skirt, socks, white footwear

  • Tracen summer uniform :

    • tracen school uniform, summer uniform, serafuku, puffy short sleeves, purple bowtie, horseshoe ornament, sailor collar, sailor shirt, purple shirt, white skirt, pleated skirt, frilled skirt, miniskirt, zettai ryouiki, white thighhighs, brown footwear, loafers

  • Tracen winter uniform :

    • tracen school uniform, winter uniform, puffy long sleeves, sailor collar, sailor shirt, purple shirt, white bowtie, horseshoe ornament, purple skirt, frills, brown footwear, loafers

Use a model derived from or mixed with animefull model.

• Settings

  • Stable Diffusion WebUI by Automatic1111

  • DPM++ Series(SDE Karras, 2M Karras, etc.)

  • about 20 steps, CFG scale 3.5~6.5

  • I recommend kl-f8-anime2.vae

  • CLIP skip = 1

  • Use Hires.fix to get higher quality image

    • Upscaler

      • Latent, Latent (nearest-exact), Latent (bicubic antialiased) or other Latent series.

    • Denoising strength 0.50~0.65

LoRA training info

• trained on sd-scripts by kohya_ss.

  • LyCORIS by KohakuBlueleaf

  • Thank you all a lot!

• based on Animefull-pruned model

• 94 images consist of

  • 4x low quality 31 images

  • 8x medium quality 32 images

  • 16x high quality 31 images

  • => 1 epoch = 876 images

• replaced character feature tags with yukino bijin \(umamusume\)

  • 1girl, blunt bangs, medium hair, horse tail, bangs, horse girl, simple background, animal ears, long hair, brown hair, diagonal bangs, official alternate costume, solo, black hair, horse ears, short hair, side ponytail, green eyes, sidelocks, yellow eyes, one side up, alternate costume, parted bangs, hair ornament, hairband, white hairband, fox tail, multicolored hair, bob cut

  • I think these tags are the essentials of Yukino Bijin or

  • unnecessary for training because of duplication.

• resolution : 768x768 with aspect ratio bucketing.

• other training settings is included in training data or metadata of LoRA file.

All uploaded images are generated by AOM3A1

And, special thanks to my best friend. DH.

In the depths of a forgotten library, amidst ancient tomes and whispered secrets, I discovered the Instrumentum insaniae—an infernal device capable of summoning Lovecraftian nightmares. Crafted from melded metals, its intricate gears and pulsating crystalline orb exuded an eerie luminescence, hinting at the forbidden knowledge it held.

With caution, I approached the artifact noticing that each dial and lever seemed poised to unlock the gateway to madness, filling me with a chilling anticipation of the cosmic horrors it could unleash. Guided by forgotten incantations, I dared to traverse the boundaries of mortal perception.

As the Instrumentum insaniae sprang to life, its eldritch radiance cast macabre shadows, transforming ordinary objects into abominable forms. Statues twisted into tentacled monstrosities, while the architecture itself shifted in alignment with the chaotic madness of the cosmos. Through its unholy power, I became a conduit, beholding the terrifying beauty of Lovecraft's cosmic pantheon.

Yet, a whispered caution echoed within my mind, warning of the perils that accompany such forbidden knowledge. With a heavy heart, I chose to part ways with the Instrumentum insaniae, leaving it behind in the hopes that others would take up its mantle, for the delicate balance that separates our world from the abyss must be preserved, and the artifact's dormant secrets may hold the key to understanding the mysteries that lie beyond.

CthulhuTECH prompt notes and fun facts

• You do NOT need to use the mythostech keyword, the lora is just that powerful, we're talking elder gods up in your GPU, some concepts will works a lot better without using it

• Check out my samples and prompts (there’s like 500 something sample images, or will be once they’re uploaded minus anything the filter rejects

• Yes, you can make porn with it, yes, some of it is confusingly hot

• If you are upscaling, keep your denoise low, you don't want to correct away some coolthulhu, With REV 1.22 and coll remacri upscaler i use 0.32 as my baseline

• Generally works fine at a strength of 1, you may want to adjust with some models

• Its ok to put 'human' in the negative, helps a lot with hybrids and deepones and if you want to do some crazy mutants

Wildcards

https://civitai.com/models/63834?modelVersionId=68415

PureChaos edition

• Purechaos edition is a lot more random, if you want to make a lovecraftian nightmare and don't care a lot about control, go forth, forsake the elder sign, invoke nyarlathotep… and adjust the strength of the lora or keyword at your peril. I’ve only included a few purechaos examples because you really don’t need them.

Glowingrunes

• glowingrunes is OP, i suggest putting it in a negative 1.0 then calling positive glowingrunes when you want them, depending on what you’re prompting a glowingrunes of 0.2 will still show over a negative 1.0 glowingrunes, especially if you’re also using cyberpunk.

So much work

• Trained on just over 3300 AI generated lovecraftian images, Generated nearly 2600 image examples curated down to just over 500.

• Coffee Maker included because this started as a konyconi style LORA, but it isn't, not anymore :D. Still they make excellent stuff and you should check em out.

• All sample images are purely this lora on top of REV 1.22, no hidden loras, TIs, tricksy editing, everything is raw prompt output.

Known Issues

• There are still a couple of weak areas that will me more cartoony, industrial equipment for example, books, and for some reason mobile phones.

• Crouching people often have 3 feet or random shoes.... I don't know why, i suspect its maybe bleed from the crouching people in robes confusing the data

• Bioluminscent doesn’t work well on anything, more of a general glow…. A lot of the training images were beautiful so i think maybe I need to train this as a general concept and not tie it to specific keywords

Leave a review

If you like it, please review it

For the love of Ithaqua buy me a coffee

https://ko-fi.com/theartificialanalyst

Prompt words

The best reference is this spreadsheet which has all the major prompts and illustrates how they fit into the hierarchy. You don’t need to use the whole hierarchy at all, but it may help to understand how things fit together (especially when working with humans, deep ones or artificial lifeforms)

The spreadsheet is currently a little incomplete because of an error I made but hopefully it’ll be completed in the next 24 hours… For now you can always get the complete list of every keyword ever by looking at the metadata with a text editor, or clicking the little ‘i’ symbol on the lora in additional networks (or via many extensions and tools)

Quickstart - if you just wanna jump in:

Human is the basic people keyword, I use man and woman for gender (you can kinda prompt futanari and easily prompt androgyny … whats the equivalent of futa for girl parts on dudes, thats like ezmode)

Mutant,Cyborg,Cultist,Demonic,Tattoo,Mask,Sorceror,Soldier,Manyeye,Dieselpunk,Fashion,Encountersuit,Cyberpunk,Scifi,Flamethrower

Monsters use Monster as their basic keyword and unless you want to be truly random I suggest actually putting in a type of monster (ghoul,chimera,manticore,unicorn etc)

Furry,Winged,Bioluminescent,Scaly,Octopoid,Humanoid,Dragon,cthulhu

Building creates buildings, I suggest being more specific e.g. a house, an apartment block, etc. You can get interiors but theyre a bit inconsistent except what ive specifically trained (e.g. cafe)

Industrial,Dieselpunk,Cyberpunk,Tendrils,Skyscraper,scaly

Technology is a catch-all for tech, its a bit mixed and needs work but if you prompt a particular thing (engine, server, air compressor) it’ll generally work ok.

Containment:Sphere,Hexahedral,pyramid - these make different shapes of containment for power plants, mystic thingoes or looking like cool AI hubs (adding eyes is cool)

Coffeemaker,pot,Mobilephone,Screen,Server,computer

Landscape c’mon ;)

Obelisks,Temple,Ruins,Pyramid,Statues,maesoleum make structures, you may want to negative them as necessary

Cyberpunk - particularly for urban or structures, dieselpunk will work a bit but it’s not trained

Jungle,Desert,bushland,Urban,Volcanic,hotspring,lava,Arctic,Underwater,abyss,forest,Park,mountain,Swamp,Hills,beach,fleshland

Hovercraft you can try some modifiers like large, small, transport but it’s inconsistent.

Cyberpunk, dieselpunk

Animal - same as monster, best to prompt a specific animal but you don’t have to.

Aquatic,Bioluminescent,Insect,serpent,

Balloon - makes zeppelins, can’t prompt much else

Cyberpunk,dieselpunk

Artifact - makes statues and small objects and things to put them in, definitely use some of the other prompts to guide it.

Statue,Obelisk,Relic,Curiosity,glowingrunes,Container,Statuette,token

Artificial - robots n such

Woman,Man,Cyborg,Android,Cyborg,Robot,Cyberpunk

Mech - Makes mechs or all types depending on how you prompt it, quads and crawlers are significantly harder to prompt

bipedal,Quad,Crawler,Biomechanical,Behemoth,Dieselpunk,cyberpunk

Deepone - lovecrafts fish-people. Seriously, check my prompts, these are hard.

Hybrid,Woman,cyberfish,Abomination,Man,pureblood,mutant,Lanternfish,goldfish,Withspawn,Angler,Bioluminescent,shark,cyborg,Encountersuit

Industrial - machinery n junk

Machine,Mining,Generator,Vehicle

Portal - y’know, portals

Doorway,rift,Primitive,Square,Circle,Hex,watcher,Tri,Scifi

Vehicle - from sports cars, to sports cars with guns, to tanks

Battlecar,Warmachine,battletank,Armed,Spikes,Spines,legs,cyberpunk,Closecombat,dieselpunk

Spaceship - sigh you understand. Can be a bit inconsistent across types of ship, prone to white background

Leviathan,Naval,Ufo,angular,Cyberpunk,Dieselpunk

Aircraft - … suggest fighter, modern fighter, bomber, etc

Antique,Cyberpunk,dieselpunk,drone

Book - On its own makes pretty cool books, doesn’t play nice with making different kinds of books.

Future plans

• Assuming people like it enough, review it, etc. I will release future versions will will include some or all of the following (if people dont like it, i'll just keep it to myself :D ) I have currently put approximately a full week of work days into this project, that's a lot of time. Making some of these changes will take considerably MORE time.
  

• Separate the major keywords into full concepts to remove some of the mixing (e.g. deepone would become a keyword not just a prompt) This is a major project in its own as i'd need to move a lot of things around (e.g. compiling all the dieselpunk and cyberpunk stuff into a keyword where it's currently in a sub sub folders

• Make separate Loras for those keywords

• Even if i don't do full concepts I will be totally reorganising the data as the way I have done it is super flawed and makes you need to prompt/negative hard if you want a very specific result

  
  

• Add more training data to some weak areas (such as industrial equipment and adding more types of books to books, aircraft, major deepone overhaul,etc,etc)

• Make a 'lite' version that is more about style than being comprehensive

• Make a 'cute' version

• Train a model (people would need to Looooooove this for me to train a model as it takes forever, at least I can run dreambooth on runpod even though that costs a few bucks)

  
  

I have a YouTube channel where I neglect to review Loras, checkpoints etc.

https://www.youtube.com/channel/UCjXlRo9kik1YJYONdjz4kiw

But i do plan to start doing it again now this ridiculous project is released.

p.s. If you 'cant get my results with my prompts that's a you problem and I can't help you.

p.p.s Cthulhu loves you anyway, dream sweetly of it.