thanks baker

File: tmp2n7ynlkq.png (409 KB, 512x640)

409 KB PNG

All my XL loras stopped working. Every output looks like this now. 1.5 loras still work. I've already made a clean install (automatic1111), downgraded to earlier versions (1.7), tried vladmandic, but nothing fixed the issue. Any thoughts? The terminal gives me no errors. It loads the lora normally, the output is just fucked.

File: chen_smile.png (1.52 MB, 1024x1248)

1.52 MB PNG

Thanks for baking

File: ComfyUI_temp_kdlef_00140_.jpg (3.42 MB, 1344x1728)

3.42 MB JPG

today i favor chacol ebara mix over autism, at least for the purposes of my deep shrink tinkering

File: fooks6.jpg (1.08 MB, 1792x2304)

1.08 MB JPG

fresh bread
>>100166445
>>100166460
yur welcome

File: UI_00134_.png (2.93 MB, 1376x1376)

2.93 MB PNG

>>100166087
I do
>>100166173
Thanks. Doesn't come out great as I would hope.

File: chun_woodenswordf.jpg (1.43 MB, 1536x2560)

1.43 MB JPG

>>100166453
That is very peculiar. If even reinstalling the UI didn't help and trying others too all I can think is to download some lora off civitai fresh and see if your local copies got corrupted somehow.

>>100166426
Prompt or catbox for OP's pic?

File: fig15.jpg (99 KB, 896x1152)

99 KB JPG

>>100166453
Yep, that makes no sense to me, especially not if you reinstalled the whole thing including venv.

Figures you could try something not auto1111 derivative for now to see if it doesn't work somewhere?

>>100166465
It also looks nice.

File: ComfyUI_temp_kdlef_00141_.jpg (3.66 MB, 1728x1344)

3.66 MB JPG

https://files.catbox.moe/664mv3.png

a few steps of the euler_smea_dy sampler was instrumental in bringing much-needed coherence to the overall gen, while not really adding much in terms of compute time because it's only for 6 steps out of 30

>>100166453
Clip skip -2 (ComfyUI) should fix it.

>>100166544
>>100166538
could the fact that I'm running it on an ext4 drive cause this somehow?

>>100166538
cute gen

>>100166561
I've tried comfy before but I could never figure out the reason why generating images through it takes >literally< 10x longer than on Auto. I tried troubleshooting but could never find the cause.

>>100166543
Not OP but here is what prompt extractor pulled.


nazrin, 1girl, solo, looking_at_viewer, short_hair, bangs, shirt, red_eyes, long_sleeves, hat, bow, holding, animal_ears, hair_between_eyes, closed_mouth, tail, white_shirt, upper_body, grey_hair, bowtie, vest, red_bow, v-shaped_eyebrows, capelet, red_bowtie, :<, +_+, brown_jacket, mouse_ears, brown_headwear, mouse_tail, magnifying_glass, deerstalker, detective

File: fig16.jpg (93 KB, 896x1152)

93 KB JPG

>>100166584
No, that seems very unlikely to me.

>>100166610
Look for A14 equivalent to
>[clip skip]
set it to -2

I have a LORA I made and really like but I wanted to try adding some more images to the data set for clarity etc.
How should I go about doing it? Can I just load the completed checkpoint and train on the dataset I want to add or should I start from the beginning again with the new data included? Or, should I start from the complete checkpoint and train using the new data and the old data?

I'd try them all myself, but I'd like to hear your thoughts before I go ahead and waste hours testing which method is best.

File: swordswoman_0007f.jpg (1.65 MB, 2048x2048)

1.65 MB JPG

>>100166594
thanks
>>100166584
Nah ext4 is not an issue. You could try checking to see if you can browse the information on the lora in the UI (click in the "i" information icon in the lora section and make sure it looks corrrect). Other than that, make absolutely sure you aren't accidentally setting it to <lora:100.0> or something instead of 1.0. And try something like comfy for sanity's sake

File: ComfyUI_temp_kdlef_00145_.jpg (3.4 MB, 1728x1344)

3.4 MB JPG

>>100166559
oh, and i bumped end_percent from 0.35 to 0.5, though idk if this is necessarily a net plus

overall you will still encounter more AI-isms like bad fingers, but for raw gens this is still leaps of improvement

>>100166695
Start from scratch and include more images to the data set.
is your best bet to do it.

if you don't believe me, just load completed checkpoint and train on the dataset you want to add.

>>100166559
>>100166702
I was excited when smea dyn was released but I haven't actually gotten the chance to try it out because of TCD autism. might give it a go later

>>100166699
guess I'll give comfy another shot then

File: ComfyUI_14361_.png (1.45 MB, 1024x1024)

1.45 MB PNG

RetroPixie based merge i'm working on, not upscaled, just genned like this

File: UI_00031_.png (1.9 MB, 1024x1024)

1.9 MB PNG

>>100166755
I'm sure the images you have from A14 can be imported in to ComfyUI so you don't lose your ....workflow I guess? I"m not sure, never used A14 not enough buttons to press.

>no debo news
He actually got hit by a bus

>>100166811
>You should go outside as well

>>100166782
v. nice.

File: dewi_00116_.png (2.41 MB, 1920x1152)

2.41 MB PNG

>>100166811
I've been too busy all day to curate the news. won't have time until later tonight. then I'm out of town thurs/fri and prob won't get a chance to do news then either :(

File: 00076-3.png (1.01 MB, 1024x712)

1.01 MB PNG

>>100166879
~woo!

thank GOD at least we get a small respite

File: ComfyUI_14367_.png (1.48 MB, 1024x1024)

1.48 MB PNG

>>100166777
trained a 1024px Lora with 285 nice background images without people in them and merged it in also, triggered by the existing "detailed background" tag (plus image-specific descriptrors and stuff), helps quite a bit I think

File: PW_67140_.jpg (281 KB, 2048x1152)

281 KB JPG

Good evening, anons! I hope everyone is doing well :]

File: UI_00114_.png (1.16 MB, 1024x1024)

1.16 MB PNG

>>100166755
Let me know if you need some workflows for comfyui so you don't feel like you've wasted your time with A14

File: 01410-In the sky Adorable(...).png (1.71 MB, 1536x1024)

1.71 MB PNG

Quokka hug balloon!

File: awoo_commits_tax_fraud.png (1.6 MB, 1024x1248)

1.6 MB PNG

1.5 models generate some nice output, but getting them to not generate extra ears = impossible

File: 00044-2766675372.png (1.15 MB, 896x1152)

1.15 MB PNG

You there! Hand over your prompts!

(or at least some cool words to use, please?)

>>100167161
white thighhighs, zettai ryouiki

>>100166941
nice

>>100167161
[land|sky|ocean]

File: PW_66996_.jpg (255 KB, 2048x1152)

255 KB JPG

>>100167118
Cute! Heya, Quokkanon :]
>>100167161
Heyy!!
(casting spell, summoning monster, portal:1.2)
Tried this yesterday and got some stupid little floating squid thing LOL

File: 01413-In the sky Adorable(...).png (1.97 MB, 1536x1024)

1.97 MB PNG

>>100167221
Hello PW

>>100167161
cohesium
powerful token, mhm

File: 00042-4169034059.png (1.19 MB, 896x1152)

1.19 MB PNG

I'll collect all the graciously shared words and do something with them later today.

>>100167221
Hey there!

Is it just me, or is training on the base SDXL model a waste of time? It seems that no lora based on it translates over well to the newer checkpoints floating around.

>>100167161
'candid' is my favorite prompt voodoo

File: UI_00065_.png (1.5 MB, 1024x1024)

1.5 MB PNG

>>100167161
(Ink illustration on parchment). A desolate wasteland stretching to the horizon, littered with remnants of advanced technology. Eerie, glowing flora sprouting amidst the wreckage. A river of molten metal cutting through the landscape, casting a vibrant red glow. Silhouettes of mutated creatures prowling in the distance. Strange symbols etched in glowing ink on the ruins. seductive. eroticism. ink splashes, rough ink sketch, moody, gloomy, dark picture, high contrast, stylized.

>>100167320
you rarely need fine tunes anyway but ones trained on the base model work fine with finetunes

>>100167202
thanks
>>100167161
style of Jamie Baldridge, style of Igor Morski

Completely missed A24 using ai to make the posters for that Civil War movie. Might be the most polished I've seen from a major prod company.

>>100167161
>Hand over your prompts!
[ style by Rob Gonsalves | style by Dan McPharlin | style by Igor Morski | style by Andrzej Grenda style by Agostino Arrivabene ]

in a pinch

>>100167161
fox ears, animal ears, animal ear fluff

apply to every prompt

File: 00029-655027935.png (2.54 MB, 1152x1920)

2.54 MB PNG

I'm not sure what mixing all of those words at once will yield, but science demands it.

I'll probably do a few other pics with just some of them to get more coherent results. Could be a good chance to try the realist models, since some of those words might not work on animagine.

File: 0.jpg (404 KB, 1024x1024)

404 KB JPG

>>100167471
>apply to every prompt
It works.

File: PW_67076_.jpg (260 KB, 2048x1152)

260 KB JPG

>>100167533
Can't wait to see what you get haha
I gotta run for a bit, but i'll be back in a few hrs :]
Btw, I finally got Pony to do backgrounds pretty well! Just adjusted some settings in PAG and it started working :]

File: ComfyUI_temp_kdlef_00184_.jpg (3.98 MB, 1728x1344)

3.98 MB JPG

https://files.catbox.moe/zvdg16.png
now i have it rigged for the last step of the euler_smea_dy pass to not downscale the model, then it resumes the downscale for the remainder to 50% of the gen

File: ComfyUI_04815_.png (3.62 MB, 2312x1920)

3.62 MB PNG

I forgot how fun this framework is. I should be able to gen my first image tonight if I just go autistic but I should make a debugger first before anything

>>100167572
hello PW! good to see you! comfy and I are going to cali again next month and we would like to come visit you!

>>100167598
it does pretty well on the details for sure. thanks for sharing your experimentation anon!

File: ComfyUI_61_.png (1.74 MB, 1024x1536)

1.74 MB PNG

Aggressive misuse of latent upscale.

>>100167622
framework?

File: 00015-1909204171.png (1.05 MB, 1104x1248)

1.05 MB PNG

>>100167326
prompt + fist of the northstar style

File: 000000_12077_.png (620 KB, 1016x680)

620 KB PNG

G'evenin Anons,

>>100166695
This for sure: >>100166717
The biggest reason IMO being that yes, your new stuff will get integrated, but you'll need to train it to do so, and you're now going to be training it against mostly the same dataset as before.
TLDR while you get your new content to 30% relevancy in the model, everything else will now be at 130%. You will overtrain.

>>100167714
do you do lora or prompt for the dress?

File: 00025-903387654.png (1.99 MB, 1024x1536)

1.99 MB PNG

>>100167728
gnevenin

File: 01414-Big antenna on head(...).png (1.14 MB, 1024x1024)

1.14 MB PNG

>>100167728
Evening anon

>>100167715
here's a hint. to make a GUI object you write:

if(NS::Button("Button"))
    counter++

funny stuff

>>100166695
using upscaled crops from the starting dataset too. you can have three images from one full body image if you crop with upper body and portrait in mind

>>100167767
>>100167772
Evenin! Nice gens

File: 00010-3570244290.png (1.21 MB, 800x1200)

1.21 MB PNG

>>100166426

we cant stop winning bros

>>100167726
I'll try it but I don't think it will come out the way you think it will.

>>100167791
groovy gen, space cat

File: UI_00143_.png (2.49 MB, 1376x1376)

2.49 MB PNG

>>100167726
my workflows not really setup for this.

File: UI_00144_.png (2.73 MB, 1376x1376)

2.73 MB PNG

>>100167726
>>100167890
I'll do one more, and going to move on to other things.

File: 00143-BMP.png (1.63 MB, 768x1200)

1.63 MB PNG

>>100166426
Blessed OP
>>100166958
Cool
>>100167883
Nice creative censorship. Also I vaguely remember Michael Jackson being in this for some reason

File: ComfyUI_ranma_04.png (1.7 MB, 1024x1536)

1.7 MB PNG

>>100167756
The dress is prompted, so it's pretty random

>>100167904
thats because he was, Space Michael is real

File: UI_00146_.png (2.34 MB, 1376x1376)

2.34 MB PNG

>>100167890
>>100167901
>>100167726

File: 0.jpg (498 KB, 1528x1024)

498 KB JPG

>>100167326

>>100167909
care to drop a prompt for the dress please?

File: 00009-2285034207.png (1.03 MB, 1104x1248)

1.03 MB PNG

>>100167890
That's pretty cool though.
>>100167928
I really like this one

>>100167942
2>7>7 CFG
>>100167928
2>3>2 CFG
Go figure.

File: The image captures a sere(...).jpg (152 KB, 1024x1024)

152 KB JPG

Afternoon Sigma bros

>>100167936
What's the CFG on this one?

File: ComfyUI_ranma_09.png (1.72 MB, 1024x1536)

1.72 MB PNG

>>100167938
covered nipples, side-tie underwear, evening gown:china dress, impossible clothes, impossible dress, bare legs, navel indent, ornate, intricate, floral pattern, skindentation
Everything related or at least affecting the dress. Start in an anime model.

Additionally, I know comfyui does : | etc differently, but the prompt is kind of just pasted from my a1111 styles while I experiment.

File: 5.jpg (610 KB, 1536x2160)

610 KB JPG

>>100167904
gotta get creative with it or you feel you're missing

>>100167978
Steps: 20, Sampler: DPM++ 2M, Schedule type: Karras, CFG scale: 7

File: 00687-1126756385.jpg (2 MB, 1344x1736)

2 MB JPG

hi bros, what image viewer you use?

File: 00012-280132880.png (1.31 MB, 1104x1248)

1.31 MB PNG

>>100167996
Is there a pony lora for it?

File: 00422-4146499890.png (2.06 MB, 1080x1536)

2.06 MB PNG

>>100168018
sxiv

>>100168027
there is, it's very attached to the style it was trained on though

File: UI_00149_.png (2.15 MB, 1376x1376)

2.15 MB PNG

>>100168003
Going to drop next image in catbox with your mix

>>100168018
allusion is the best ive found
>>100168027
nice

Anyone got tips on making non-pony loras work a bit better with pony? especially photorealistic?

>>100168045
just train one

>>100168049
against pony as the base model? prolly dumb question. have had shit luck training against anything other than base XL

>>100168061
guides are floating around for training realistic on it. why not give those a read?

cargo cult thread

File: 00013-280132881.png (1.34 MB, 1104x1248)

1.34 MB PNG

>>100168044
Thanks, adding fist of the northstar to stuff, wasn't sure how much I liked her in the style. also check.

>>100166717
>>100167746
Thanks, guess It's back to training from scratch again.
>using upscaled crops from the starting dataset too
I may just do this. The new upscale tile models for SDXL should help and there are some fine details I want the model to capture.

File: 00697-357545496.jpg (2.04 MB, 1344x1736)

2.04 MB JPG

>>100168030
>>100168044
thanks

>>100168074
10-4 thanks

File: UI_00150_.png (2.1 MB, 1376x1376)

2.1 MB PNG

>>100167928
>>100167936
>>100167942
>>100168003
https://files.catbox.moe/lxrxuo.png
catbox is my image.

Your settings didn't come out right.

File: 00096-3642368582.png (1.83 MB, 1024x1536)

1.83 MB PNG

>>100168018
Honeyview

>>100168116
catbox is kill for some reason. I'll try again in a bit

Did you hear that NVIDIA discovered magic numbers that make your generation quality better?

SDXL: 14.615, 6.315, 3.771, 2.181, 1.342, 0.862, 0.555, 0.380, 0.234, 0.113, 0.029

Stable Diffusion 1.5: 14.615, 6.475, 3.861, 2.697, 1.886, 1.396, 0.963, 0.652, 0.399, 0.152, 0.029

File: UI_00151_.png (2.25 MB, 1376x1376)

2.25 MB PNG

>>100168138
Oh, this is the image that it got.

>>100168143
https://www.youtube.com/watch?v=vVPT0JT1dOw

>>100168143
where do I put them?

>>100168143
>1girlposting in paper
They know.

>>100168164
you got to zoom out a bit.

File: ComfyUI_ranma_14.png (1.69 MB, 1024x1536)

1.69 MB PNG

>>100168159
https://research.nvidia.com/labs/toronto-ai/AlignYourSteps/

>>100168160
https://files.catbox.moe/ndn4ka.json

>>100168173
outpainting is a good idea. there are a few other things i need to polish as well.

>>100168177
but I use euler.

or what they simply did is what we all do.
XY plot CFG for each check point until you get magic numbers for it that look great.

File: UI_00123_.png (1.21 MB, 1024x1024)

1.21 MB PNG

>>100168196
litterbox is dead.

File: BMP_10070_.png (767 KB, 496x1328)

767 KB PNG

>>100167996
True. Damn I miss all the wacky ps2 rhythm games so much.

>>100168241
That was on Dreamcast

>>100168251
Must have been ported over to ps2 then because I never had a dreamcast

>>100168261
SEGA died and thus everyone started porting all their games to other platforms.

>>100168237
>>100168160

Okay, fine, I'll link to reddit then.
https://www.reddit.com/r/StableDiffusion/comments/1ccdt3x/nvidia_presents_align_your_steps_workflow_in_the/

File: UI_00156_.png (2.53 MB, 1376x1376)

2.53 MB PNG

>>100168319
Cheers m8.

It saves me 2 seconds on my image gens.

>>100168143
They really want the consumer/gamer audience to pick up on SD so there is more preference for their RTX over AMD cards. Wouldn't be surprised if they include a section about Stable Diffusion performance in their 5090 reveal live stream.

>>100166453
Try changing the prompt a bit. If you have weights, change them up to more than .5

File: Screenshot.jpg (1.37 MB, 5639x5622)

1.37 MB JPG

>>100168319
I don't use comfy that much but was going to play around with this. I'm missing a couple nodes. When I go to manager then click install missing nodes I just get a blank window and nothing happens

>>100168496
Got to sell overprized shit somehow since AMD caught up to them and cheaper. The only way they can sell their crap is "WE HAVE CUDA! crunch your AI numbers with our graphics cards because AMD doesn't have it" That's why AMD released their MES as open source just to catch up on any aspect.

>>100168532
>that gay ass prompt

its true that ai enthusiasts are just big perverts

File: Capture.png (50 KB, 2028x190)

50 KB PNG

>>100168532
Left one is Seed generator. Just use any seeder and connect it the same way.

The other one is Sigma generator, I wish I knew how to pull it up in my manager so I can give you the exact wording to search for.

File: ComfyUI_00035_.jpg (1.05 MB, 2560x1440)

1.05 MB JPG

>>100168542
Sometimes I think I'm the only ai enthusiast who doesn't generate coom. Why are 95% of ai bros degenerate anime gooner incels?

Every single AI story falls into three categories
1) Evil AI destroys world
2) Good AI is basically God
3) You can use a computer to avoid dealing with real women, but it might not always work the way you hope
And you're suprised people use the infinite beautiful woman machine to make infinite beautiful women? When the AI can fulfil scenarios 1 or 2, we can regroup.

>>100168616
3 heavily realize on 2.
2 cannot exist without 3.
1 is waiting for 2, and 3 to start working so it will win by default.

>>100168599
hey, some of us do artistic landscapes AND smut

>>100168647
that 1% AI generated landscape folder oooh baby!

>>100168578
I found those 2. I've been having that problem with manager for a while where it just gives me a blank window when I click install missing nodes

File: Reeeeetsu.png (994 KB, 896x948)

994 KB PNG

>>100168616
Jokes on you, been doing #3 for years without the use of AI

>>100168664
stop attacking my lifestyle

>>100168698
Do you know how to pull up what node name it is?

>>100168755
what is sigma under in nodes?

>>100168698
>>100168755
try KJNodes for ComfyUI

>>100168780
I installed it and restarted and it seems that was the missing node

>>100168809
Nice, good luck tweaking it.

>>100168599
The problem with AI and coom is that all coomable generally one of the strongest indicators of a well trained model. If AI wasn't good for copming, it wouldn't be good for much else because by being bad at cooming it has failed a fundamental step in human interaction.

>>100168820
Libido and art were always intertwined. The 1001 Nights nerds (the equivalent of their era) made up a perfect fellow-nerd waifu to engage in the otherwise culturally-stigmatized act of storytelling, exactly like a modern anime fan creating a girl who's into insert-topic-here. The Greeks imagined art as a beautiful female spirit interacting with a male artist. None of what's happening is new, not even the puritains furious that other people are enjoying what they personally consider sinful.

File: ComfyUI_00011.png (1.19 MB, 1024x1024)

1.19 MB PNG

>>100168815
Thanks! I changed it to 1girl and got this trash gen

File: dewi_00120_.png (2.38 MB, 1920x1152)

2.38 MB PNG

will the evil Ai that destroyed humanity run in comfyui? how much vram will it need?

File: UI_00157_.png (1.61 MB, 1024x1024)

1.61 MB PNG

>>100168861
It just saves few seconds of time. Nothing special and breaks a lot of things that used to work.

>>100168820
>If AI wasn't good for copming, it wouldn't be good for much else because by being bad at cooming it has failed a fundamental step in human interaction.
Pictures of cocks, nipples, and prolapsed assholes in a dataset doesn't really help generate pictures of automobiles anon... I'm not sure you understand how image diffusion models work.

>add a single feature to workflow
>give no thoughts to organization and readability
>work out kinks
now i gotta go back through and reorganize everything goddamnit

>>100168820
>If AI wasn't good for copming, it wouldn't be good for much else
You don't remember everyone saying XL was only good for non coom?

File: UI_00163_.png (1.92 MB, 1024x1024)

1.92 MB PNG

>>100168928
it takes 100 tries but you'll get there.

File: 86698-564797748-score_9, (...).png (2.12 MB, 1536x1536)

2.12 MB PNG

>>100168879
What style LoRA?

File: ComfyUI_04523_.png (3.82 MB, 1920x2312)

3.82 MB PNG

>>100169010
masakasama, it's my own bake but I think I have to get rid of the clean linework so I can get the textured stuff more

File: 86134-4183697074-score_9,(...).png (507 KB, 960x657)

507 KB PNG

>>100169049
Love anything that separates it self from the "AI norm"

File: UI_00192_.png (2.69 MB, 1248x1824)

2.69 MB PNG

>>100169003

File: dewi_00121_.png (2.42 MB, 1920x1152)

2.42 MB PNG

>high-resolution spatially-varying
bidirectional reflectance distribution functions (HRSVBRDFs)
fuck these people and all their bullshit

File: UI_00195_.png (2.91 MB, 1248x1824)

2.91 MB PNG

>>100169003
>>100169064
“Magic Mirror on the wall, who is the fairest of them all?”

Painfully dumb question but XL is like the newer, more advanced version that needs higher hardware, am I getting it right? For example for Dreamshaper there is 8 and the XL, and if your hardware can handle it, there is no point not to go for XL? FP16/32 is similar and one can just go for higher one if the rig can handle it, right?

comfyfags with thicc vram, have you guys played with this at all? https://civitai.com/models/319652

File: dZ1623031703.jpg (161 KB, 1416x1024)

161 KB JPG

>>100169110
Yes

File: UI_00145_.png (1.4 MB, 832x1216)

1.4 MB PNG

>>100169138
>https://civitai.com/models/319652
Nope, because not even previews look all that great.

File: 00010-2221455619-86af7495(...).png (3.15 MB, 1432x1840)

3.15 MB PNG

>>100169064
>>100169096
:]

File: dewi_00127_.png (2.17 MB, 1920x1152)

2.17 MB PNG

>>100169110
>FP16/32 is similar and one can just go for higher one if the rig can handle it, right?
this is the opposite. if your card can run lower precision natively, it results in much faster inference with no quality loss

>>100169144
>>100169174
Oke, thank you, guys.

File: UI_00147_.png (2.95 MB, 1248x1824)

2.95 MB PNG

>>100169168

File: 00015-2221455624-86af7495(...).png (3.19 MB, 1432x1840)

3.19 MB PNG

>>100169209

File: 00041-1341062857-((1girl)(...).png (2.85 MB, 1920x1088)

2.85 MB PNG

>>100167936
Full prompt/settings for this one? Looks awesome

File: ComfyUI_04515_.png (3.97 MB, 1920x2312)

3.97 MB PNG

>>100169138
>ultranationalism /
ComfyUI_Pangu_Draw_V3
nope

Dalle-3 won

>>100169278
It was nerfed to the point of no return by early November

File: BMP_10078_.png (2.05 MB, 1328x1328)

2.05 MB PNG

>wanted to make PtR earlier
>keeps making UjL in Parappa clothing
SD3-SAN SAVE MEEEE

File: UI_00164_.png (2.84 MB, 1248x1824)

2.84 MB PNG

>>100169214

>>100168599
>pats self on the back

>>100168599
There used to be more like you... But there were always still more coomers

File: 00035-3222035510-baa42631(...).png (3.14 MB, 1840x1432)

3.14 MB PNG

>>100169278
Congrats, take your trophy

File: 00029-4183917855-09790486(...).png (2.91 MB, 1840x1432)

2.91 MB PNG

>>100169291

File: UI_00180_.png (2.73 MB, 1248x1824)

2.73 MB PNG

>>100169301
>>100169357
I've reloaded to anime style

File: 00027-4279580617.png (1001 KB, 1024x1024)

1001 KB PNG

I have all your prompts now. Time to cook.

File: 00035-4159137092.png (1.29 MB, 1104x1248)

1.29 MB PNG

>>100169368
Nice, the realism was looking really good though

File: 00034-4159137091.png (1.37 MB, 1104x1248)

1.37 MB PNG

>>100169368
>>100169384
But I also like that

File: UI_00184_.png (2.58 MB, 1248x1824)

2.58 MB PNG

>>100169357
>>100169368
>>100169384
Bed time calls though. We'll continue the venture in 13 hours.

can i not used tiled samplers with masks?

>>100169368
She looks deliciously evil

File: dewi_00131_.png (2.54 MB, 1920x1152)

2.54 MB PNG

>>100169377
youll never take me prompts

>>100169377
You can have mine.

>vampire queen, black royal dress, black royal miniskirt, bare legs, sitting on a throne, red tiara, naughty, romantic, white hair, long hair, simple background, mist, bats, (from front), parted lips, subtle smile, beautiful eyes, naughty,

Continue my battle after I'm gone.>>100169400

File: 00028-4279580618.png (1.2 MB, 1024x1024)

1.2 MB PNG

Thread is ending so I'll begin posting stuff in the new one.

>>100169406
I'll just add some horns in there.

>>100169421
I'll add that to the next batches, the cooking is in progress.

File: DEBO_00055_.png (2.1 MB, 1728x1344)

2.1 MB PNG

>>100169446
>I'll just add some horns in there.
oh ne he figured out me prompts

File: UI_00196_.png (1.55 MB, 1024x1024)

1.55 MB PNG

>>100169225
>(Ink illustration on parchment). A desolate wasteland stretching to the horizon, littered with remnants of advanced technology. Eerie, glowing flora sprouting amidst the wreckage. A river of molten metal cutting through the landscape, casting a vibrant red glow. Silhouettes of mutated creatures prowling in the distance. Strange symbols etched in glowing ink on the ruins. seductive. eroticism. ink splashes, rough ink sketch, moody, gloomy, dark picture, high contrast, stylized.

with >>100168319
workflow. Good luck

File: UI_00197_.png (3.21 MB, 1536x1536)

3.21 MB PNG

>>100169490
I would post catbox but it is down at the moment

File: 0.jpg (145 KB, 1024x1024)

145 KB JPG

put "little man, distant man" in negs

File: 00032-655027938.png (2.51 MB, 1152x1920)

2.51 MB PNG

I should have gone horizontal for the first batch.

New
>>100169489
>>100169489
>>100169489
Thread

>>100169377
fug

>>100169490
NTA but I'll take this for a spin

File: 00029-2649553324-(((playg(...).png (3.72 MB, 1920x1088)

3.72 MB PNG

>>100169490
Thanks buddy ole pal

>>100169400
gn