Previous /sdg/ thread : >>101343417>Beginner UI local installFooocus: https://github.com/lllyasviel/fooocusEasyDiffusion: https://easydiffusion.github.ioMetastable: https://metastable.studio>Local installAutomatic1111: https://github.com/automatic1111/stable-diffusion-webuiComfyUI: https://github.com/comfyanonymous/ComfyUIAMD GPU: https://rentry.org/sdg-link#amd-gpuIntel GPU: https://rentry.org/sdg-link#intel-gpu>Use a VAE if your images look washed outhttps://rentry.org/sdvae>Auto1111 forksSD.Next: https://github.com/vladmandic/automaticAnapnoe UX: https://github.com/anapnoe/stable-diffusion-webui-ux>Run cloud hosted instancehttps://rentry.org/sdg-link#run-cloud-hosted-instance>SD3 info & downloadhttps://rentry.org/sdg-link#sd3https://education.civitai.com/quickstart-guide-to-stable-diffusion-3https://aitracker.art/viewtopic.php?t=57>Try online without registrationsd3: https://huggingface.co/spaces/stabilityai/stable-diffusion-3-mediumtxt2img: https://www.mage.spaceimg2img: https://huggingface.co/spaces/huggingface/diffuse-the-rest>Models, LoRAs & upscalinghttps://civitai.comhttps://huggingface.cohttps://openmodeldb.info>Animationhttps://rentry.org/AnimAnonhttps://rentry.org/AnimAnon-AnimDiffhttps://rentry.org/AnimAnon-Deforum >Index of guides and other toolshttps://rentry.org/sdg-linkhttps://rentry.org/rentrysd>View and submit GPU performance datahttps://vladmandic.github.io/sd-extension-system-info/pages/benchmark.htmlhttps://docs.getgrist.com/3mjouqRSdkBY/sdperformance>Share image prompt info4chan removes prompt info from images, share them with the following guide/site...https://rentry.org/hdgcbhttps://catbox.moe>Discord6wUwtcJsr2>Related boards>>>/h/hdg>>>/e/edg>>>/d/ddg>>>/b/degen>>>/vt/vtai>>>/aco/sdg>>>/trash/sdg
>>101349367
I fixed the issue. Python 3.12 is not compatible with any version of SD.
>>101349510don't bother upgrading python for at least a year, they basically release beta versions every few months with breaking changes. 3.13 is out in a couple months lol
Its been a while since i generate images. Is sd3 supported by automatic webui? Why is top searches show that it doesnt
>>101349699Why are you searching instead of checking the repo itself lol
i cannot get the fingers fixed. Any good loras? Maybe also for teeth?
I once was a kidAll I had was a dream
>>101349916what changed?
>>101349925Love and disease and wisdom
>>101349321will any of the cloud hosts let me use Loras?
>>101349987Paperspace
>>101349916More like: Damn kids, it's not even a beach anymore.
>>101350235Nope
>>101347190>"It looks like Jin"Not wrong, the vaguely fashion photography look was Jin's bread and butter. Does he ever still post though? I haven't seen him a while.
I don't get how everyone seem to have their pony working well, but the face's of mine look like this I have been using StabilityMatrix rather than A1111 (because it randomly stops working).
>mfw Resource news07/10/2024>ConceptExpress: Harnessing Diffusion Models for Single-image Unsupervised Concept Extraction https://github.com/haoosz/ConceptExpress>EasyAnimate-V3-XLhttps://huggingface.co/alibaba-pai/EasyAnimateV3-XL-2-InP-960x960>AMD to Acquire Silo AIhttps://www.globenewswire.com/news-release/2024/07/10/2911168/0/en/AMD-to-Acquire-Silo-AI-to-Expand-Enterprise-AI-Solutions-Globally.html07/09/2024>Tailor3D: Customized 3D Assets Editing and Generation with Dual-Side Imageshttps://tailor3d-2024.github.io>PaintsUndo: A Base Model of Drawing Behaviors in Digital Paintingshttps://lllyasviel.github.io/pages/paints_undo>Video-STaR: Self-Training Enables Video Instruction Tuning with Any Supervisionhttps://orrzohar.github.io/projects/video-star>Compositional Video Generation as Flow Equalizationhttps://adamdad.github.io/vico>Big Tech needs to generate $600 billion to justify AI hardware expenditurehttps://www.techspot.com/news/103699-big-tech-needs-generate-600-billion-annual-revenue.html>Minutes to Seconds: Speeded-up DDPM-based Image Inpainting with Coarse-to-Fine Samplinghttps://github.com/linghuyuhangyuan/M2S>GenArtist: Multimodal LLM as an Agent for Unified Image Generation and Editinghttps://zhenyuw16.github.io/GenArtist_page>AID-AppEAL: Automatic Image Dataset and Algorithm for Content Appeal Enhancement and Assessment Labelinghttps://github.com/SherryXTChen/AID-Appeal>Align and Merge Deep Learning Modelshttps://github.com/shoroi/align-n-merge>UltraEdit: Instruction-based Fine-Grained Image Editing at Scalehttps://ultra-editing.github.io07/08/2024>JeDi: Joint-image Diffusion Models for Finetuning-free Personalized T2I Generationhttps://research.nvidia.com/labs/dir/jedi>PartCraft: Crafting Creative Objects by Partshttps://github.com/kamwoh/partcraft>AnySR: Realizing Image Super-Resolution as Any-Scale, Any-Resourcehttps://github.com/CrispyFeSo4/AnySR
>mfw Research news07/10/2024>V-VIPE: Variational View Invariant Pose Embeddinghttps://v-vipe.github.io/>Latent Space Imaginghttps://arxiv.org/abs/2407.07052>HumanRefiner: Benchmarking Abnormal Human Generation and Refining with Coarse-to-fine Pose-Reversible Guidancehttps://arxiv.org/abs/2407.06937>Rethinking Image-to-Video Adaptation: An Object-centric Perspectivehttps://arxiv.org/abs/2407.06871>Beyond Aesthetics: Cultural Competence in Text-to-Image Modelshttps://arxiv.org/abs/2407.06863>Chat Edit 3D: Interactive 3D Scene Editing via Text Promptshttps://sk-fun.fun/CE3D/>Powerful and Flexible: Personalized Text-to-Image Generation via Reinforcement Learninghttps://arxiv.org/abs/2407.06642>Mobius: An High Efficient Spatial-Temporal Parallel Training Paradigm for T2V Generation Taskhttps://arxiv.org/abs/2407.06617>Iteratively Refined Image Reconstruction with Learned Attentive Regularizershttps://arxiv.org/abs/2407.06608>Vision language models are blindhttps://arxiv.org/abs/2407.06581>Sketch-Guided Scene Image Generationhttps://arxiv.org/abs/2407.06469>VIMI: Grounding Video Generation through Multi-modal Instructionhttps://arxiv.org/abs/2407.06304>A Survey on Mixture of Expertshttps://arxiv.org/abs/2407.06204>Poetry2Image: An Iterative Correction Framework for Images Generated from Chinese Classical Poetryhttps://arxiv.org/abs/2407.06196>More Distinctively Black and Feminine Faces Lead to Increased Stereotyping in Vision-Language Modelshttps://arxiv.org/abs/2407.06194>VideoEval: Comprehensive Benchmark Suite for Low-Cost Evaluation of Video Foundation Modelhttps://arxiv.org/abs/2407.06491>MiraData: A Large-Scale Video Dataset with Long Durations and Structured Captionshttps://arxiv.org/abs/2407.06358>EMBANet: A Flexible Efffcient Multi-branch Attention Networkhttps://arxiv.org/abs/2407.05418>Learning to Adapt Category Consistent Meta-Feature of CLIP for Few-Shot Classificationhttps://arxiv.org/abs/2407.05647
news you (i) can, but wont, use
>>101351190>MiraData: A Large-Scale Video Dataset with Long Durations and Structured CaptionsSurely there's a better way of detecting scene changes than using 2 (two) models with billions of parameters each>black frame? probably a cut>next frame is slightly different than the last? panningDoes this even need AI
>>101345939
>>101351289its a lot more challenging to detect the boundaries are of events within scenes, which is also most relevant for video captioning. where does a car chase start or stop, where are people talking to each other, when was a coffee cup picked up, etc. these boundaries are important for both t2v and vqa
>>101349510Works fine with ComfyUI. I have been using python 3.12 for over 6 months now.
Does this look hand made?
>>101351727looks like a.i.
>>101351727no because of how wobbly the finer lines are. a human artist would usually try to have a more consistent structure
>>101351747>>101351769but will the average normie realize is AI?
>>101351787no
>>101351787they are more likely to tell that it is a.i. than to believe it is 'hand made'
>>101351787unlikely. we're double biased, first because you're posting to an AI image thread but also because were already used to looking at and picking out AI images. normies probably wouldn't think twice about it
>>101351821>>101351809>>101351808I want to make a coloring book.
>>101351787It is literal visual noise with zero aesthetic redemption the average normie wont consider it at all for good reason
>>101351836you'd need to iterate on and refine your process a bit more but its perfectly achievable
What should the base model of loras I use with aDetailer be? The same one as the model for the whole image, or SD 1.5 or something?
>>101351836whether that is viable, with that quality of generation, would depend on what is selling currently. seeing how successful shamelessly slop a.i. stuff is on facebook etc, who knows.
>>101352004nvm, weird but I tried a few celebrity loras for faces in pony realism and only one worked
>>101351727>Does this look hand made? 100% AI>>101351787>but will the average normie realize is AI?I am an average normie. If in doubt I say it's AI. My normie friends do the same.
https://youtu.be/YW6eHunEysA?si=dLV9Bt3sp_U8eyty
https://github.com/lllyasviel/Paints-UNDO/issues/19kek, why is Illya getting spammed by butthurt "artists" lmao
>>101352742What are we listening to here? I think ill stay with Dave Brubeck - Take Two .. Bake Two seems to be abit to baked for my taste. Not even the slimes are amused.
>>101352763this is not a legitimate use for the github issue systemcan you not just auto close these things?
>>101352763it works good for generic low effort stuff like anime 1girls, but fails with complex scenes, vector designs and photographic realism https://lllyasviel.github.io/pages/paints_undo/
>>101352926illy is a master troll. the whole point of his work is to generate this kind of content
>>101353107Hi ani
Morning anons
>>101353170Gmoorning! Looks delish
c89.5>TD0JA0Dear DOJ, my robot says you are a joke."Said."
>>101353411are you using a new lora? this style is very fun
>>101353411>>101353194imposter green hair anon
I have a requestI'm using stealth png info, it's meant for discord trannies but I think it works for 4chin toocan you guys drop picrel into png info in your sdg instance?Does it show the metadata? It works for me
>>101353516no
>>101353537debo please stop seething about le pisscord
>>101353516>stealth png info4chan only scrubs exif, not alpha channel. So there's no reason why it wouldn't work
>>101353516sounds sus
>>101353516>https://github.com/neggles/sd-webui-stealth-pnginfoLink to the thing if you too are interested in embedding the metadata in the image pixelsThis way you won't be jewish like >>101353537>>101353554Le pisscord stopped stripping metadata from images, but I wonder if this still works for 4chin
>>101353554not debo
>>101353516it does not work
>>101353596Can you please install this repo as an extension and try again?>https://github.com/neggles/sd-webui-stealth-pnginfo
>>101353487Green hair anon? my my, I ain't that .. just made a space ship captain girl with a crew of slimes
>>101353607forgot you gen
>>101353597noob question: is CLIP powerful enough to retrieve prompts used to generate a picture, like this one for example? How to best retrieve the prompt used for this pic?
>>101353626
>>101353712you're describing the 'interrogate' mechanism but, no, it can't retrieve a prompt from interrogating the image. it can give some ballpark concepts but it'll be very far off from whatever the initial input was. you'd get closer using an LLM to describe/caption the image than using interrogate
It works! hurrayA shame you need to install the repo, I thought it would've worked without, thank you.Now we need to make that extension mandatory in these threads.
>>101351787Before AI would assume you did a really shitty Photoshop filter.
What planet should the slime crew visit? Gimme ideas!
>>101353808Nice
>>101353713Thank you anon
>>101353835see if you can get them visiting an alien marketplace/promenade, checking out all the alien wares
>>101353820I like your work
>>101353848noprob
>>101353597>>101353610it does work, it's hiding the data in the LSB of every pixel, so like if you have a fully red pixel [0b11111111, 0b00000000, 0b00000000] you can hide three bits in it, if you wanted to hide [0b010] you'd end up with [0b11111110, 0b00000001, 0b00000000] which still *looks* like a pure red pixel to a humanuseful but i like the social aspect of asking people for catboxes
>>101353610dumb
>>101353951It's a dumb premise because you're not entitled to people's generation information.
>>101353951oh great so soon we'll have executeables from opening other people's catboxes, what could go wrong
>>101353857Yea! Old theme I did once I really love.. its abit challenging.. reminds me of a certain music video to.
Hey traniLeave those kids alone!
>>101354000Its a jewish* premise, You're a jew (and you should kys)
>>101354000checked, also are you retarded? baka at least own the fact that you're using a model trained on all of the internet"noooo my sacred prompt you can't have it" so fucking cringe>>101354034are you also retarded? that's not how any of this works, this doesn't open up any additional vulnerability, read the extension code
>>101354039hardest part is getting the 'alien' quality of the alien marketplace since 'marketplace' is such a biased token for terrestrial bazaars and the like
she got way more alien, but I kinda like it
>>101354090yea "alien" bled onto the character.. but I guess thats okay.. maybe she disguised cause aliens are humanophobe
I can think of only one known jew ITT
>>101354069Only jews stick their hands in other people's pockets>>101354083Only low IQ retards need to steal other people's work and you're why I'm never helpful and why I love training custom models I never share.
>>101354161>love training custom models I never share.Sure, anon. Sure.
>>101354172I believe her
>>101354161deep irony from a cringe nogen
>>101354172Not everyone is as poor and stupid as you. You act like training a lora or model is particularly difficult lmao. At least admit you're under 18.
>>101354130>inferior anon
>>101354190guys look how good this 1girl isat least I get why you try to steal prompts because this is what you do on your own
>>101354099more alien more better
nigga stole my prompt
>>101354241If you take something that isn't yours without permission that is stealing. It's also really stupid in an era where more and more fragmentation goes on. What, you too stupid to write with your imagination?
>>101354229but its against the lore! .. "alien" is a very strong tag in ponx/autismmix I guess, I replaced it with science fiction and it kinda works.. also I forgot "nude" in negatives first .. thanks porn model! https://files.catbox.moe/eb8ryn.png
Somehow that anon is far more of a nuisance than debo.
>>101354090Is it me or does this look like mikhaila petersonIs there a mikhaila peterson lora for pony?
>>101354507>mikhaila petersondidn't know her, had to google.. no.. doesnt look like her at all
>>101354266>science fiction and it kinda works
>>101354507extremely vaguely barely not really when you ask me to look really hard
>>101354000dont use it then retard
>>101354645The whole thing relies on you on tricking people into using it :)
>>101354605lovely!
>>101354921
>>101354933>you have been violated girly man
>>101354954
>>101354976hey guys i'm trans get it
>>101354995i don't get it
>>101355091Your parents failed you
>>101355122i suspect we have that in common.
>>101355154I don't post on /b/
>>101355196nice. id pat you on the head, or wherever else you'd prefer, if able
>>101355224Your life sucks
Does anyone here use marijuana? I just started on it for a medical condition and I'm looking for a dry cape to avoid smoking if anyone has any recs
weird horns
>>101355413Looks like cross between a jester hat and bunny ears
>>101355321could be worse. rn im enjoying a diet dr pepper, in my air conditioned apartment, candle burning, sunny day... oh how horrible.
>>101355436Sounds mega comfy anon
>>101355411>I'm looking for a dry capehuh
>>101355445Vape sorry.
i miss schizo anon
>>101355506I still have no idea who this is
>>101355411I'd still take the smoking personally, vapes taste like chemicals and fuck up your health. Dab rigs and extracts are the way to go imo, way less harsh than bongs and don't smell.
Tifa posting is back bros hehe, a staple of sdg(pic unrelated)
>>101355553>I'd still take the smoking personallyI didnt ask>vapes taste like chemicalsHe said DRY vape>and fuck up your healthSee above>Dab rigs and extracts are the way to go imo, way less harsh than bongs and don't smellNot what I asked
Should I do facedetailer before or after hires fix?
>>101355506This is debo posting
>>101355632wrong thread
>>101355632I've waffled back and forth on this, so I think ultimately it depends on the model(s) and workflow. its nice to end with the hires pass because that will make the detailer changes more consistent with the rest of the image. but it can also damage the detailing if your hires pass is very high denoise or just not the best with detail. on the flip side, sometimes ending with the detailer causes the touched areas to not jive with the surrounding areas.I think either setup can work if you tweak the params well. try both and see which performs more reliable to your tastes>>101355655I don't like schizo anon
>classic debo non answer
native Kolors support in comfy when?
>>101355742whats your answer?
>>101355791why would i tell you?
ITS SO HOT
>>101355922it is! or it was actually.. 34°C earlier.. than thunderstorm, now its sauna here
>>101355937You know there are different places on earth other than right where you are right
>>101355822I made a song about how smart and helpful and cool I amhttps://suno.com/song/4441743c-8cc4-4e7d-bba1-fb321812ae84
>>101355947no shit sherlock!
>>101355964This is the most cringe thing ever posted here
>>101356004heh, mad
>>101356064I'm not mad I'm just pointing out that its cringe
>>101355517newfag
>debo makes a song about himselfAnd you thought he couldn't be any more pathetic
gotta goo... time to move on again /sdg/ was fun for a while! see yas
>>101356075>I'm not madclassic mad response. here, enjoy another song I made about how great I amhttps://suno.com/song/894332a0-7bda-4915-8994-f6e9d461ee3a>>101356172bye bye, see you again
>>101356144Yeah idk the retarded lore everyone here engages in sorry I have a life bro
>>101356184see >>101356144
>>101356183Seems like projection. Weird
>>101356183why do you even post that man have some pride in what you do
>>101356244make something betteroh, you cant, lol
93f outside, 70f inside, like 33c and 21c ithink.
nb4 "stop bullying me"
>>101355947Spastic
>>101355715You doxxed yourself spamming the same message.
>>101355715>I don't like schizo anonmean.
the dream
How the hell do I resize without image quality taking a massive poopoo in cumfy?
Spastic
>>101356486Downscale or upscale? Downscaling is pretty easy. Upscale is a whole can of worms
i've seen some great multiple_views/concept art/sketches posted before, anyone got a catbox prompt they could share so i could compare notes and see what i'm doing wrong? mine always turn out chonk. please and thank you.
>>101357254Why don't you post an example image and a prompt and we can work it out...
>>101357254https://files.catbox.moe/4lo0gm.png
>>101349321when lora training a character, if i want the characters apprearance to be consistent should i have emotions and poses be used as tagged (e.g amrs up, hands on hips, smiling, blushing) when training or should i remove most of it, because I get weird results when i tag emotions and try to use them
>>101357081tried. wish i could go higher but it makes my 3080 crash. do I need a 5090 for high res upscales anons?
holy trinity.
>>101357442i keep getting wierd realistic lips on an anime character trained on an anime model and being run on an anime model, no fucking idea why since theres 27 reference photos
>>101357539did you make it yourself? did you tag the lips?
>>101354312talking shit
I didn't keep up with SD since almost a year back.Is it still mostly tag based or can you write in natural language?
>>101357467Try to download the 'ultimate sd upscale' and use it. It uses tiling so you can do it with a lot less VRAM consumption. Though it can take some practice to get the hang of its configuration options
>>101357750what tile size should i use? 512 or 1024
Please brosIs there a way to reduce the size of thumbnails?
>>Auto1111 forks>SD.Next: https://github.com/vladmandic/automatic>Anapnoe UX: https://github.com/anapnoe/stable-diffusion-webui-uxIs it worth changing from Forge to one of these?
>>101357796lora thumbnails* in sd forge
>>101357287i think my issue is that i'm on nai v3 so she's fatally allergic to subjects>1https://files.catbox.moe/7qwsek.png>>101357410so smooth, i jelly
>>101357788The options I use when using ultimate upscale, and I by no means guarantee they are best, are (for SDXL-size base resolutions like 1024x1024 and 2x upscale):Type: ChessTile Width/Height: 1024Mask Blur: 32Padding: 256 (128 in comfy as it's the highest it will go I believe)Upscaler: 4x_NMKD-Superscale-SP_178000_G for realistic or classic fantasy style illustrations, R-ESRGAN 4x+ Anime6B for anime stuff
>>101351610Cool yin yang,
>>101357742>can you write in natural language?yes
>>101357932Thanks anon, guess I'll reinstall automatic UI.Been so long updating mine would break everything lol.
>>101357965I'm using ComfyUI and scared to update sometimes, if it breaks it's usually fixed quickly though. The T5 encoders I believe do the natural language, in SD3 and kolors I know for sure.
>>101357742yes but the best models these days are pony descendants and it helps if you tag promptpeople say kolors and pixart etc are good but the best 1girls are like 80% some form of pony
>>101357838naisisters... not like this
>>101358027ComfyUI allows more stuff than automatic or it's just an esthetic/UI choice between the two?I have no idea what "T5" or "kolors" are, one year and I feel like a total newbie again...>>101358056I'm mostly interested in nice poses, no limbs looking like shit, etc.And damn it, prompt still is the best way then...
>>101358138Yes, ComfyUI is updated the quickest, kolors is a new model that ComfyUI is able to use.This image is kolors in comfyUI then sd3 ultimatesd upscaled.
>>101358138comfy lets you get really into the weeds but unless you're trying to push the envelope and spend a million hours it's not better than auto, it also has a much worse API if that's something you care aboutit's worth learning both and you can set them up to share the same models/loras etc so you really don't have to chose between themif you want specific poses just use a pose controlnet, both auto and comfy have addons that just let you pose a stickfigure and then the gen has the same pose
>>101358181I see.>>101358310Yeah I was there for controlnet, just wish I could write in natural language "x is at the left of y, and z is behind them making a peace sign" instead of tags + controlnet.
>>101358118>naisisterscalling me a dumpster fire would be a compliment
>>101358181that’s beautiful anon, thank you
>>101358479You're welcome, ty, kolors is a good model.
>>101353626is stealth png info on the automatic1111 extensions list?
>>101357563yes, but i didnt tag anything to do with lips, ive experimented with lots of tags and not very many tags, seems like my best attempt didnt use too many tags but after cleaning up the reference images i cant seem to get the same quality
>>101358699Couldn't find it in there no
Some tits for the thread
>>101358793prefer robins
>>101358699>https://github.com/neggles/sd-webui-stealth-pnginfoyou just add the url "install from url"
>>101358800uwu
ccccombo breaker
>>101359157habbo hotel meets the urbz
Either it's me or mermaids that look inhuman but still beautiful are easier said than done
done
>>101359289wrong kind, do a fish head with human legs
>>101359487shit I would drink a can of monster right there
I ended up commissioning an artist only to start getting sneaking suspicions that AI is being used. Just random minor things that look off throughout the process. If I wanted AI to do it, I'd have just learned it myself and done it for free to the exact specs I wanted. How do I even accuse the artist I commissioned? I don't want to pay the remainder when I'm getting charged hand-drawn prices.
>>101359555Larp or post example
>>101359555I don't mind as long as they're clear about it.I'd rather have a well edited AI image by an artist who knows how to do that well, than some ugly stuff.
>>101359555>I could have done it myself for free to the exact specs I wanted>um guys, do you think I'm getting a bad deal?
>>101359564I'm still in the middle, I don't need her running off to twitter and getting me cancelled.>>101359582That's the problem she's anti-ai judging from what I saw. She's probably using the hand drawn excuse to get away with charging a lot more. But with all the errors I'm spotting it feels like she fed it into an AI, gave up after a few tries, and just sent it as is with several errors like they were genned separately and then put together. I could've done a better job if I learned this AI stuff. Probably could've saved a lot if I found a good AI artist to do it, at least they'd be upfront about it.
>>101359555Post it here and we can probably find AI mannerisms. You could also try that 'is it AI' website and use the results as evidence, but it's not always correct. I have no idea what it's accuracy percentage is.
>>101359680>That's the problem she's anti-ai judging from what I saw."The loud antigay advocate was secretly gay all along!"How not surprising kek.
need to revive term 'proompting'
>>101359680Yeah too bad. You might be easy to identify because you would be one of her few active commisions. There won't be a next time, but you could've just asked here if we thought she was a fraud without revealing yourself as a client.
>>101359772thats pw's favorite word. well, I guess behind 'purple' and 'witch'
I've made 3 loras for this kid on civitai and he just asked for another one. Only this time he said 'and send it soon.' I think I'm going to have to just ignore him because he's not going to stop. Maybe people don't realize making the datasets is pretty time consuming.
>>101359973>and send it soonidk that sounds like a threat. he sounds tough. I wouldn't want to be on his bad side
>>101359973Why are you slacking? I said I wanted it soon.
>>101359973>and send it soonWhat a polite fellow.
>>101357838Is there new nai3 based models?
>>101360069I thought they never leaked.
>>101358181>upscaled>899x1572What's your base image size? Isn't kolors 1024x base?
>>101359973>send it soonWow he doesn't have long left to live. Thanks for making his wish come true
>>101360179goobular
>>101360215thangyoo
>>101360093I then shrink it down in photoshop to 4meg to post here.
last
>>101360235You can use jpg 100% compression up to 4k I think.
>>101360320>>101360320>>101360320>>101360320
>>101360038>>101360063It took me a couple weeks on the last one. I think he's actually trying to tell me not to take so long this time...
>>101350564nice
>>101359275new dwarf fortress mod lookin good