[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: 1719248797021213.png (1.84 MB, 1600x896)
1.84 MB
1.84 MB PNG
Previous /sdg/ thread : >>101130602

>SD3 info & download
https://rentry.org/sdg-link#sd3
https://education.civitai.com/quickstart-guide-to-stable-diffusion-3
https://aitracker.art/viewtopic.php?t=57

>Beginner UI local install
Fooocus: https://github.com/lllyasviel/fooocus
EasyDiffusion: https://easydiffusion.github.io
Metastable: https://metastable.studio

>Local install
Automatic1111: https://github.com/automatic1111/stable-diffusion-webui
ComfyUI: https://github.com/comfyanonymous/ComfyUI
AMD GPU: https://rentry.org/sdg-link#amd-gpu
Intel GPU: https://rentry.org/sdg-link#intel-gpu

>Use a VAE if your images look washed out
https://rentry.org/sdvae

>Auto1111 forks
SD.Next: https://github.com/vladmandic/automatic
Anapnoe UX: https://github.com/anapnoe/stable-diffusion-webui-ux

>Run cloud hosted instance
https://rentry.org/sdg-link#run-cloud-hosted-instance

>Try online without registration
sd3: https://huggingface.co/spaces/stabilityai/stable-diffusion-3-medium
txt2img: https://www.mage.space
img2img: https://huggingface.co/spaces/huggingface/diffuse-the-rest

>Models, LoRAs & upscaling
https://civitai.com
https://huggingface.co
https://openmodeldb.info

>Animation
https://rentry.org/AnimAnon
https://rentry.org/AnimAnon-AnimDiff
https://rentry.org/AnimAnon-Deforum

>Index of guides and other tools
https://rentry.org/sdg-link
https://rentry.org/rentrysd

>View and submit GPU performance data
https://vladmandic.github.io/sd-extension-system-info/pages/benchmark.html
https://docs.getgrist.com/3mjouqRSdkBY/sdperformance

>Share image prompt info
4chan removes prompt info from images, share them with the following guide/site...
https://rentry.org/hdgcb
https://catbox.moe

>Discord
6wUwtcJsr2

>Related boards
>>>/h/hdg
>>>/e/edg
>>>/d/ddg
>>>/b/degen
>>>/vt/vtai
>>>/aco/sdg
>>>/trash/sdg
>>
Cursed thread
>>
File: 00004-4265728316.png (2.63 MB, 1280x1920)
2.63 MB
2.63 MB PNG
There is actually nothing wrong with these threads
>>
File: 1698080306775921.jpg (1.15 MB, 3024x1728)
1.15 MB
1.15 MB JPG
>>
>hahahah prove it prove it lalala :]
>https://files.catbox.moe/j06k2r.JPG
>*vanishes*
i wish you would just fuck off forever trani, im not joking
>>
It doesn't look much like a Fez
>>
>>101135016
prove it :]
>>
File: da noos 7.jpg (612 KB, 1344x768)
612 KB
612 KB JPG
>mfw Resource news

06/24/2024

>Stable Cascade was originally MIT licensed?
https://github.com/Stability-AI/StableCascade/commit/209a52600f35dfe2a205daef54c0ff4068e86bc7

>Geometry-Aware Large Reconstruction Model for Efficient and High-Quality 3D Generation
https://github.com/alibaba-yuanjing-aigclab/GeoLRM

>VideoGenHub: One-stop library to standardize inference and evaluation of Conditional Video Models
https://github.com/d8ahazard/VideoGenHub

>AI’s Most Ambitious Music Generators Accused of ‘Massive’ Infringement In New Lawsuit
https://www.rollingstone.com/music/music-news/record-labels-sue-music-generators-suno-and-udio-1235042056/

>ECLIPSE: Expunging Clean-label Indiscriminate Poisons via Sparse Diffusion Purification
https://github.com/CGCL-codes/ECLIPSE

>Contextual Interaction via Primitive-based Adversarial Training for Compositional Zero-shot Learning
https://github.com/lisuyi/PBadv_czsl

06/23/2024

>MOFA-Video: Controllable Image Animation via Generative Motion Field Adaptions in Frozen Image-to-Video Diffusion Model
https://github.com/MyNiuuu/MOFA-Video

>SD.Next release 2024-06-23
https://github.com/vladmandic/automatic/blob/master/CHANGELOG.md#update-for-2024-06-23

>ComfyUI Unique3D
https://github.com/jtydhr88/ComfyUI-Unique3D

>AutoStudio: Crafting Consistent Subjects in Multi-turn Interactive Image Generation
https://github.com/donahowe/AutoStudio

06/22/2024

>ComfyUI DiffSynth wrapper nodes
https://github.com/kijai/ComfyUI-DiffSynthWrapper

>Update on SD3 on Civitai
https://civitai.com/articles/5840

>ExVideo: Post-tuning technique aimed at enhancing the capability of video generation models
https://huggingface.co/ECNU-CILab/ExVideo-SVD-128f-v1

>ComfyUI-JNodes: Variety of python and web UX improvements
https://github.com/JaredTherriault/ComfyUI-JNodes

>StableSwarmUI will no longer be maintained under Stability AI
https://github.com/Stability-AI/StableSwarmUI#migration-notice
>>
>mfw Research news

06/24/2024

>Image Conductor: Precision Control for Interactive Video Synthesis
https://liyaowei-stu.github.io/project/ImageConductor/

>HLQ: Fast and Efficient Backpropagation via Hadamard Low-rank Quantization
https://arxiv.org/abs/2406.15102

>A3D: Does Diffusion Dream about 3D Alignment?
https://arxiv.org/abs/2406.15020

>Disability Representations: Finding Biases in Automatic Image Generation
https://arxiv.org/abs/2406.14993

>VividDreamer: Towards High-Fidelity and Efficient Text-to-3D Generation
https://narcissusex.github.io/VividDreamer/

>Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning
https://arxiv.org/abs/2406.15334

>LLM2FEA: Discover Novel Designs with Generative Evolutionary Multitasking
https://arxiv.org/abs/2406.14917

>Demonstrating the Efficacy of Kolmogorov-Arnold Networks in Vision Tasks
https://arxiv.org/abs/2406.14916

>LatentExplainer: Explaining Latent Representations in Deep Generative Models with Multi-modal Foundation Models
https://arxiv.org/abs/2406.14862

>Six-CD: Benchmarking Concept Removals for Benign Text-to-image Diffusion Models
https://arxiv.org/abs/2406.14855

>Is A Picture Worth A Thousand Words? Delving Into Spatial Reasoning for Vision Language Models
https://arxiv.org/abs/2406.14852

>Latent diffusion models for parameterization and data assimilation of facies-based geomodels
https://arxiv.org/abs/2406.14815

>Evaluating Numerical Reasoning in Text-to-Image Models
https://arxiv.org/abs/2406.14774

>Holistic Evaluation for Interleaved Text-and-Image Generation
https://arxiv.org/abs/2406.14643

>Stylebreeder: Exploring and Democratizing Artistic Styles through Text-to-Image Models
https://stylebreeder.github.io/

>DragPoser: Motion Reconstruction from Variable Sparse Tracking Signals via Latent Space Optimization
https://arxiv.org/abs/2406.14567

>This Looks Better than That: Better Interpretable Models with ProtoPNeXt
https://arxiv.org/abs/2406.14675
>>
>>101135067
ok >>98185914 (search on moe or any of the other archives)
>>
/sdg/ is back.
>>
oh god please not another pedo cope thread
>>
>>101134652
I take it back, epicRealisimXL is really fucking good
>>
File: ComfyUI_01389_.png (3.67 MB, 2389x1536)
3.67 MB
3.67 MB PNG
>>
>Trani in shambles
>>
>ani gets exposed by multiple anons
>ran no longer mentioned
>>
Any illustrators that img2img their shit for references here? That thing is a godsend. Just started using ponyxl and its taught me so much, someone request a character or something and ill sketch it out then img2img as an example
>>
I am shocked, shocked to find that there is a pervert in /sdg/. Anyway,
>>
if i join the discord am i going to be groomed by trani?
>>
>>101135258
you will, for sure. probably not anyone else tho
>>
>>101135258
if you don't have any self control then no. If you do then still no.
>>
>>101135258
no he just searches for anons gooning with him, he is harmless otherwise
>>
>>101135000
If you've lurked here for years, you can pinpoint the many times the quality plummeted. The thread becoming the discord's designated shitting street is just the next step in this circus.

this is all because of the anons that allowed it to happen, from the ones who thought they could manipulate the thread with their samefagging and sock puppets. To the ones that tried to turn sdg into their daily blog. Those anons opened up space for bad actors to turn sdg into a discord grooming thread.
>>
>>101135214
ran is still worse than ani. nothing changed
>>
>>101135289
every video game general degenerates over time as normal people lose interest. I'm sure the Skyrim generals are garbage in much the same way this thread is
>>
>>101135316
Elaborate
>>
>>101135316
Distilled cope
>>
Unironically, I think Ani is a ticking time bomb
>>
>>101135330
ran is obsessed and must mention debo and ani in everything he says
>>
>>
>>101135342
this
>>
Please leave your discord drama on the discord, guys. I promise you most of us don't think about "ani" at all, you don't need to warn me about some guy I can already smell through my screen on the rare occasion that he actually posts anything.
>>
>trani is running away like always
pathetic
>>
File: rider1.png (3.67 MB, 1965x1306)
3.67 MB
3.67 MB PNG
>>
File: 1345151346162456.jpg (182 KB, 640x868)
182 KB
182 KB JPG
>>
>>101135469
based and true
>>
>>101135469
kek
>>
julien is shit
>>
>>101135143
i can't recreate https://civitai.com/images/11700743 strangely
>>
File: 1716698201200168.jpg (1022 KB, 3024x1728)
1022 KB
1022 KB JPG
>>
will schizo anon ever be satisfied?
>>
>>101135218
i do it the other way, txt2img and use that as reference
>>
so this thread is for SAI shilling, julien generating pictures of naked children and being mad at "chinks"? did i get that right?
>>
>>101135707
*and the discord
>>
>Julien is crying
>>
>>101135819
don't know who that is
>>
>>101135835
Ani/fizzledorf
>>
>>101135871
prove it :]
>>
its been like 6 months since ive been here. whats all the new shit ppl are using. do ppl still not know how to upscale 1.5s? iss SD3 shit?

clean install btw so either A1111 or comfy or whatever might be better now
>>
>>101135934
go to /ldg/ or join the discord instead. this is a shit posting general now
>>
>>101135908

>>101135097
>>
File: 1711357464467232.jpg (1.29 MB, 1432x1840)
1.29 MB
1.29 MB JPG
>>101135934
same answer as last thread, SDXL models are real good, Comfy is powerful and modular, PonyRealism is my current favorite. SD3 is fun but I am the best prompter around so YMMV
>>
>>101136023
catbox link is dead for me
>>
>>101135959
how about post gens and stop trying to start shit like a feminie little bitch. are you able to do that oh wait, too late. if you need dick so badly bend over
>>
File: sd3bo_00022_.png (3.53 MB, 1728x1024)
3.53 MB
3.53 MB PNG
>>101135934
>whats all the new shit ppl are using
everyones using different things depending on their sensibilities and goals. hard to give a blanket answer to this
>do ppl still not know how to upscale 1.5s?
not sure I understand, people have been upscaling for forever. do you mean something more specific?
>iss SD3 shit?
yes
>clean install btw so either A1111 or comfy or whatever might be better now
comfy has a lot more dev support now. the a1111 dropped off pretty hard so they're a lot slower to add new things in or update stuff to latest versions.
>>
File: prod.jpg (97 KB, 974x1083)
97 KB
97 KB JPG
I keep reading Prodigy is better and I keep trying it but the results are not great, so i'll ask. what do I set all of this to? Clearly I don't know what i'm doing.
>>
File: ComfyUI_temp_ryvar_00005_.png (2.48 MB, 1536x1536)
2.48 MB
2.48 MB PNG
>>
>>101135934
SD3 is weird. It doesn't diffuse the same way, and that has lots of unpredictable effects. The dataset tagging is also different, so it can be hard to figure out how you're supposed to prompt it. To give one example of how SD3 is weird, saying too strongly that a woman has a big nose can cause her nose to turn into weird visual artefact garbage—you can watch this happen usually in the last half of diffusion steps. It can be very difficult to get the right prompt for an SD3 image, but rewarding when you succeed.

In the end I think they really crippled it with how they AI-captioned training data. And it's undertrained compared with XL. Worth trying but overall a big disappointment, even with my modest expectations.
>>
>>101136025
ur gen is dookie

if you cant 1 red brick, git gud
>>
File: ComfyUI_temp_ryvar_00008_.png (2.82 MB, 1536x1536)
2.82 MB
2.82 MB PNG
>>
>>101136105
>saying too strongly that a woman has a big nose can cause her nose to turn into weird visual artefact garbage
sounds like sd3 will be better for AI video
>>
>trani still a pedo
>>
File: ComfyUI_temp_drfoc_00030_.png (2.66 MB, 1536x1536)
2.66 MB
2.66 MB PNG
>>
>im not a pedo you dum dum
>its just "shota"
>>
File: 1708669975407906.jpg (1.52 MB, 3024x1728)
1.52 MB
1.52 MB JPG
>>
>>101136212
you've been at this for hours
>>
>>101136223
never forgetti :]
>>
>>101136234
who?
>>
>>101136251
trani
>>
>>101136257
prove it :]
>>
File: grin.jpg (242 KB, 1024x1024)
242 KB
242 KB JPG
>>
File: ComfyUI_01424_1.png (3.23 MB, 2190x1408)
3.23 MB
3.23 MB PNG
>>
>>101136262
20k lineart lora lmao :]
>>
did ani post at all today?
>>
>>101136309
Yes, you missed the posts.
>>
>>101136219
control net?
>>
>>101136112
lol welcome back 1 red brick anon
>>
File: 00200-3101635887.png (2.11 MB, 1024x1656)
2.11 MB
2.11 MB PNG
>>
>>101136314
prove it :]
>>
>>101136319
thanks homie
>>
>>101136330
You need to be more available. It's not my fault.
>>
i miss schizo anon
>>
>>101136334
>>101136368
also me
>>
>>101136348
I can't see him anon, where is ani? :]
>>
File: 00059-3046431651.png (1.12 MB, 1024x1024)
1.12 MB
1.12 MB PNG
>>
>consuming or creating loli or any derivative
>differentiating between pedo and hebe
If this is you, seek help
>>
File: 00027-2567702229.png (1.12 MB, 1024x1024)
1.12 MB
1.12 MB PNG
>>101136413
Agreed
>inb4 pedo apologists.
>>
File: ComfyUI_temp_ijjnm_00009_.png (3.54 MB, 1997x1331)
3.54 MB
3.54 MB PNG
>>
>>101136413
w-why are you still l-looking at it anon... you wouldn't happen to be... a pedophile?
>>
>>101136413
true
>>
>trani trying to defend pedo positions
you know i did like you besides you being a sperg
but i will not defend a pedo so i hope you die in a fire trani
>>
I keep looking at shota ani's asshole. it really does look like a quivering bunny nose
>>
File: fefeghhdd.jpg (184 KB, 832x1216)
184 KB
184 KB JPG
>>
File: 00426-2284758804.png (1.17 MB, 1024x1024)
1.17 MB
1.17 MB PNG
>>
>>101136464
i said this a long time ago and dipped. all those avatar fags are pedos. i called comfy out on it constantly
>>
i miss schizo anon
>>
File: 1696612628676803.jpg (1.24 MB, 3024x1728)
1.24 MB
1.24 MB JPG
>>
>>101136484
what does shota comfy's asshole look like :]
>>
>>101136515
glazed doughnut.
>>
File: 1714630815772133.jpg (1.23 MB, 3024x1728)
1.23 MB
1.23 MB JPG
>>
>meltdown over multiple days
he got fired again didnt he kek
>>
File: 00022-2634501384.png (967 KB, 1024x1024)
967 KB
967 KB PNG
>>
>>101136530
kek
>>
File: 1714941054377307.jpg (1.49 MB, 2880x1624)
1.49 MB
1.49 MB JPG
>>
>official SAI employee
>>
File: 1715290603032053.jpg (1.44 MB, 3024x1728)
1.44 MB
1.44 MB JPG
>>
File: 1704596306995723.png (744 KB, 925x1384)
744 KB
744 KB PNG
Is there a workflow to turn a sobel picture made from a photobashed collage into an inked page without color or that tries to mimic inked pages?
>>
File: 00039-2634501401.png (1.06 MB, 1024x1024)
1.06 MB
1.06 MB PNG
>>
File: 1699363926950588.jpg (1.66 MB, 3024x1728)
1.66 MB
1.66 MB JPG
>>
>Wow, did he really just..? Wow. Drawing distinctions? Defending the bad thing even in a very limited context or heavily-qualified sense? That's... not a good look! You just told on yourself my man. Caught in 4K. This one's going in the archives. You could have simply touched grass and had a normal one but instead you chose to die on this hill. That's, uh, pretty gross. Enjoy your call from the FBi dude. Play stupid games, win stupid prizes.
>>
File: sd3bo_00154_.png (3.14 MB, 1728x1024)
3.14 MB
3.14 MB PNG
>>101136589
you can maybe use a controlnet like canny or scribble
>>
File: ghffhfhhfhf.jpg (163 KB, 832x1216)
163 KB
163 KB JPG
>>
>>101136589
Hey I know that cat! Isn't he the one with short lefs on youtube.
>>
>>101136602
hot
>>
62k+ dataset with 1024+ resolution buckets takes just over 24 hours for one epoch, if I do 100 epochs the model will be in a reasonable state late October
can someone recommend another hobby
>>
>>
>>101136684
crocheting
>>
File: home.png (3.04 MB, 1536x1536)
3.04 MB
3.04 MB PNG
decided to do something "artistic" instead of my usual taylor edits
>>
>>101136631
but how do I prompt it to be monochrome inked page?
>>
>>101136701
Looks fascinating. I might gen something not so sure yet.
>>
are all SAI employees p****?
>>
>>101136688
this reeks of ancestral sampling
>>
File: grehhfefeh.jpg (179 KB, 832x1216)
179 KB
179 KB JPG
>>
>>101136728
Not at all. It's a diverse team from all over the world. But I can see how you'd think so given that the main face of SAI was Bangladeshi
>>
>>101136728
yes, me included
>>
File: sd3bo_00152_.png (2.59 MB, 1728x1024)
2.59 MB
2.59 MB PNG
>>101136716
the controlnet extracts the edges in preprocessing then your prompt is applied to the output in sampling. you'd be prompting something like "monochrome inked page"
>>
File: 1695685105663435.jpg (1.73 MB, 3024x1728)
1.73 MB
1.73 MB JPG
>>
File: 1713474129248392.jpg (1.53 MB, 3024x1728)
1.53 MB
1.53 MB JPG
>>
File: 1.jpg (174 KB, 768x1280)
174 KB
174 KB JPG
>>
Is there a way to import lora training metadata into one of the lora trainers? I'm trying to do it manually but it's very confusing.
>>
>>101136716
You need to separate the meaning
>in black and white ink,
maybe
>printed comic book page,
You can get away by img2img'ing the original.
To reinforce the style you can add 'by some known black and white ink artist'. I don't know any out of my head right now.
>>
File: ComfyUI_temp_drfoc_00097_.png (2.03 MB, 1536x1536)
2.03 MB
2.03 MB PNG
>>
cooking for red brick anon as we speak. I AM the best prompter.
>>
File: 1699069834381378.jpg (1.29 MB, 3024x1728)
1.29 MB
1.29 MB JPG
>>
I have all the proof in the world ani
>>
File: 00196-TFT_1249571.png (3.61 MB, 1382x2304)
3.61 MB
3.61 MB PNG
>>
File: 1691816540473112.jpg (1.65 MB, 3024x1728)
1.65 MB
1.65 MB JPG
>>
File: greta.jpg (86 KB, 1152x832)
86 KB
86 KB JPG
>>101136716
>portrait of Greta Thunberg in black and white ink, (ink painting:1.2)
AlbedoXL but I am sure vanilla is more than fine.
>>
i can assure you it will only get worse from this point onwards :]
>>
>>101136734
interesting how low the cfg can go with the Euler SMEA / AyS scheduler
>>
File: yjjkyu.jpg (376 KB, 832x1216)
376 KB
376 KB JPG
>>101136878
nice atmosphere.
>>
>>101136975
what? the smell? take a shower then
>>
File: 1696158771407141.jpg (826 KB, 3024x1728)
826 KB
826 KB JPG
>>
>>101137011
you will not need to worry about this once im finished with you :]
>>
>>101136929
>portrait of Greta Thunberg in black and white ink, (ink painting:1.2)
>>
File: ComfyUI_temp_drfoc_00110_.png (2.63 MB, 1536x1536)
2.63 MB
2.63 MB PNG
>>
Amazing the amount of time I spend inpainting and editing lewd images no one will ever see but me
>>
>>101137054
clownussy hits different
>>
>>101137051
Well I don't how to force strict ligne claire type of art. Maybe need to use a specific artist or maybe it also understands 'ligne claire'.
>>
File: 1692565421277166.jpg (1.21 MB, 3024x1728)
1.21 MB
1.21 MB JPG
>>
File: 1716449731287004.png (3.74 MB, 1536x1536)
3.74 MB
3.74 MB PNG
i'm bricked up
>>
File: 1708460099915017.jpg (1.42 MB, 3024x1728)
1.42 MB
1.42 MB JPG
>>
>>101136871
none that I could find I'm afraid, surprising considering you can import .json files
would have though extracting would have been an easier process
>>
>>101137035
please be gentle
>>
File: ComfyUI_temp_zhvqy_00003_.png (3.65 MB, 1992x1328)
3.65 MB
3.65 MB PNG
>>
File: 1718964329179677.png (2.53 MB, 1536x1536)
2.53 MB
2.53 MB PNG
it's a brick kind of day today
>>
File: 1689674929511061.jpg (1.35 MB, 3024x1728)
1.35 MB
1.35 MB JPG
>>
File: gffss.jpg (146 KB, 832x1216)
146 KB
146 KB JPG
>>101137024
noice
>>
File: greta2.jpg (119 KB, 1152x832)
119 KB
119 KB JPG
>>101137051
Here's another.
>portrait of Greta Thunberg in black and white ink in comic book page, comic book layout, (ink painting:1.2)
>>
>newfags taking advice from debo

You know he’s elated
>>
File: 1714957926955501.jpg (1.45 MB, 3024x1728)
1.45 MB
1.45 MB JPG
>>
>>101137100
I can assure you I'm not him. I am posting here because I want to post and I don't give a fuck about your drama one way or another.
>>
File: ComfyUI_01450_1.png (3.9 MB, 2389x1536)
3.9 MB
3.9 MB PNG
>>
>>101137112
I wasn’t talking about you
>>
File: 1692992847816578.jpg (1.27 MB, 3024x1728)
1.27 MB
1.27 MB JPG
>>
File: 00118-1018449433.png (1.37 MB, 1024x1024)
1.37 MB
1.37 MB PNG
>>
>>101137129
Maybe stop being so passive aggressive then.
>>
File: 1707631298135946.jpg (1.7 MB, 3024x1728)
1.7 MB
1.7 MB JPG
>>
File: 1696976787857020.jpg (1.19 MB, 3024x1728)
1.19 MB
1.19 MB JPG
>>
File: 07069-2778292655.png (1.36 MB, 1344x768)
1.36 MB
1.36 MB PNG
>>
god i need to shit
>>
File: 1688491274482731.jpg (765 KB, 3024x1728)
765 KB
765 KB JPG
>>
File: 00052.jpg (227 KB, 1088x1560)
227 KB
227 KB JPG
>ask for bald woman, get what I asked for
idk my minds eye has something closer to sinead o'connor
>>
File: 1698006202875975.jpg (1.08 MB, 3024x1728)
1.08 MB
1.08 MB JPG
>>
File: 000000_13848_.png (2.62 MB, 963x1571)
2.62 MB
2.62 MB PNG
>>101137091
Nice.

>>101135078
debs,, hopefully Matteo gets it working with sd3 soon.. https://github.com/cubiq/ComfyUI_IPAdapter_plus
https://www.youtube.com/watch?v=ewKM7uCRPUg
>>
File: 1705855958715471.jpg (1.1 MB, 3024x1728)
1.1 MB
1.1 MB JPG
>>
File: 1690795459746622.png (3.9 MB, 1536x1536)
3.9 MB
3.9 MB PNG
>>
File: 00302-3039118343.png (1.39 MB, 1024x1024)
1.39 MB
1.39 MB PNG
>>
File: 1702210122980459.jpg (1.99 MB, 3024x1728)
1.99 MB
1.99 MB JPG
>>
FUUUUUCK bros I miss having proper stable diffusion so much. I used to trialscum on gravitai but recently found out they locked it behind a subscription. Any recs?
>>
Best hi-res fix anime style upscalers? Any tips on avoiding hands getting ruined during these hi-res fixes?
>>
File: 1696508346924055.png (3.72 MB, 1536x1536)
3.72 MB
3.72 MB PNG
>>
File: 00204-3039118335.png (1.32 MB, 1024x1024)
1.32 MB
1.32 MB PNG
>>
File: 1701493114659834.jpg (2.14 MB, 3024x1728)
2.14 MB
2.14 MB JPG
>>
File: 00195-327723629.jpg (74 KB, 1216x832)
74 KB
74 KB JPG
>>
File: 1703443594893967.jpg (1.62 MB, 3024x1728)
1.62 MB
1.62 MB JPG
>>
File: grggddd.jpg (125 KB, 1216x832)
125 KB
125 KB JPG
Best waifu?
>>
File: 00135-3039118336.png (1.3 MB, 1024x1024)
1.3 MB
1.3 MB PNG
>>
>>101135078
>AI’s Most Ambitious Music Generators Accused of ‘Massive’ Infringement In New Lawsuit
It's over
>>
File: 1697688159241794.jpg (1.41 MB, 3024x1728)
1.41 MB
1.41 MB JPG
>>
File: 1691737986702208.jpg (1.31 MB, 3024x1728)
1.31 MB
1.31 MB JPG
>>
File: greta5.jpg (67 KB, 1152x832)
67 KB
67 KB JPG
>portrait of Greta Thunberg, (by Hergé:1.2), in the drawing style of ligne claire
By Herge creates an association about the general colors and the fact she's wearing a blue woolen shirt.
Maybe TCD lora is affecting this too.
>>
File: 1688814004269593.jpg (1.33 MB, 3024x1728)
1.33 MB
1.33 MB JPG
>>
>>101137301
fat old hag
>>
Can you do regional prompt easily with drawthings (OSX)?
>>
>>101137332
Not the point here.
>>
File: 1694681981364033.jpg (985 KB, 3024x1728)
985 KB
985 KB JPG
>>
>>101137269
Chang'e in the back
>>
>>101137345
Me btw
>>
File: 1718050412099541.jpg (2.03 MB, 3024x1728)
2.03 MB
2.03 MB JPG
>>
File: 00044-3388670845.png (1.3 MB, 1024x1024)
1.3 MB
1.3 MB PNG
>>101137290
What model is that?
>>
File: greta6.jpg (163 KB, 1152x832)
163 KB
163 KB JPG
>>101137301
Vanilla is much better at this than AlbedoXL.
At least I would call this ligne claire...
>>
>discord
>>
>imagine thinking drawn shota is real
>>
File: 1717646992679468.jpg (1.46 MB, 3024x1728)
1.46 MB
1.46 MB JPG
>>101137360
diamondvisualxl_v20
>>
File: 1709042925011838.jpg (1.75 MB, 3024x1728)
1.75 MB
1.75 MB JPG
>>
>>101137409
I don't think anyone is confusing real versus not. It's just a bad look for anyone to be airing that out
>>
File: sd3bo_00023_.png (2.79 MB, 1728x1024)
2.79 MB
2.79 MB PNG
>>101137216
>https://github.com/cubiq/ComfyUI_IPAdapter_plus
is this pre-breaking news? I'm not sure what to do with this
>>
>>101137301
(Alfred E. Neumann)
>>
>>101137409
>flaccid strawman
>>
File: 1716950844150614.jpg (1.75 MB, 3024x1728)
1.75 MB
1.75 MB JPG
>>
File: 1710526482936079.jpg (1.82 MB, 3024x1728)
1.82 MB
1.82 MB JPG
>>
>>101137348
>Chang'e
great choice
>>
File: 1710347944477452.jpg (1.23 MB, 3024x1728)
1.23 MB
1.23 MB JPG
>>
File: 1706451152913544.jpg (1.5 MB, 3024x1728)
1.5 MB
1.5 MB JPG
>>
>>101137466
>is this pre-breaking news? I'm not sure what to do with this

dope node added to his pack,, thought it deserved news!
>>
File: 00076-3388670850.png (1.34 MB, 1024x1024)
1.34 MB
1.34 MB PNG
What's the verdict on SD3?
>>
File: 1708914079413463.jpg (1.74 MB, 3024x1728)
1.74 MB
1.74 MB JPG
>>101137539
Bad
>>
File: 1697195072175859.jpg (1.28 MB, 3024x1728)
1.28 MB
1.28 MB JPG
>>
>>101137552
can you please stop flooding these threads
>>
>>101137552
>Bad
too bad,
I'm not finding diamondvisualxl_v20 on hugging face or civit
>>
>doxxing yourself to own the haters
>>
File: 1707621839069571.jpg (1.31 MB, 3024x1728)
1.31 MB
1.31 MB JPG
>>101137562
This is you
>>
File: bedroom.jpg (225 KB, 1152x832)
225 KB
225 KB JPG
>>101137562
>>
File: 07090-4091638345.png (1.17 MB, 1344x768)
1.17 MB
1.17 MB PNG
>>
File: 1696362604466612.jpg (1.32 MB, 3024x1728)
1.32 MB
1.32 MB JPG
>>101137573
Try just searching for diamondvisual - probably it's called 2.0 or etc.
>>
File: 00059-3388670851.png (1.35 MB, 1024x1024)
1.35 MB
1.35 MB PNG
>>101137562
no
>>
is training on 1000+ pictures stupid?
>>
File: 1702336915392214.jpg (897 KB, 3024x1728)
897 KB
897 KB JPG
>>
>>101137601
you have to train on 0 images
>>
File: 1693744339747847.jpg (1.56 MB, 3024x1728)
1.56 MB
1.56 MB JPG
>>
>>101137601
depends on your goal
>>
File: 1699574702196101.jpg (1.2 MB, 3024x1728)
1.2 MB
1.2 MB JPG
>>
File: 1695846688973833.jpg (841 KB, 3024x1728)
841 KB
841 KB JPG
>>
>>101137573
You're being sent on a snipe hunt, FYI
>>
File: 1688868614030070.jpg (1.11 MB, 3024x1728)
1.11 MB
1.11 MB JPG
>>101137562
This is also you
>>
File: 1693765651525065.jpg (1.4 MB, 3024x1728)
1.4 MB
1.4 MB JPG
>>
Apologize to ran
>>
>>101137576
it was one, of many, once in a lifetime opportunities they had to pursue
>>
>>
File: 00007-2108808723.png (2.78 MB, 1280x1920)
2.78 MB
2.78 MB PNG
>>101137675
Who should
>>
>>101137649
oh, why would someone do that?

>>101137596
>You're being sent on a snipe hunt, FYI
Why would you do this to me? I thought we were cool.
>>
I failed...
>>
>>101137687
I will save your image for another gallery image in the near future.
>>
I'll gen something else now.
>>
File: 1702520863223635.jpg (1.05 MB, 3024x1728)
1.05 MB
1.05 MB JPG
>>101137697
Well, I guess it's no longer on civit, but here's a review of it from 8 months ago for us to enjoy together
civitai com/posts/794390
>>
File: 00012-2704317.png (2.85 MB, 1280x1920)
2.85 MB
2.85 MB PNG
>>101137706
I will be on the look out for it
>>
>>101137624
can you explain? i'm training one person with many outfits and trying to get their body down.
>>
File: 1699500181375984.jpg (1.53 MB, 3024x1728)
1.53 MB
1.53 MB JPG
>>
autismmixconfetti on SDXL... how do I prompt for the bra being lifted off over the boobs? I can't seem to get it to work. Shirt lifting, yes - bra lifting, no.
>>
File: 1709419348807901.jpg (1.47 MB, 3024x1728)
1.47 MB
1.47 MB JPG
>>
>>101137725
It is outside my current attention but I will do it at some point and you are going to know.
>>
So i recently got a folder with 640 images of Carrie Fisher and 99 of those images are high-res, is that overkill for a LoRA?
>>
File: 1695469013759860.jpg (950 KB, 3024x1728)
950 KB
950 KB JPG
>>
File: 1707188451948682.jpg (1 MB, 3024x1728)
1 MB
1 MB JPG
>>
>>101137719
ah yes, I have a few models that got this treatment as well.
>>
File: 1719273084907_image.png (2.02 MB, 1024x1536)
2.02 MB
2.02 MB PNG
Wrong season
>>
File: 1700126346942884.jpg (1.15 MB, 3024x1728)
1.15 MB
1.15 MB JPG
>>
if anyone is interested here is using a lora vs base pony model (mixes) for 2B, chosen because 2B is in extremely high amounts of data in the base model
>>
File: ComfyUI_temp_inuzp_00023_.jpg (394 KB, 1200x1800)
394 KB
394 KB JPG
>>
>>101137773
who's that?
>>
>>101137757
Of Carrie Fisher, yes.
>>
File: ComfyUI_temp_jpzno_00002_.png (1.73 MB, 1120x1440)
1.73 MB
1.73 MB PNG
>>
>>101137825
Is SDXL fine-tunes better at realism? Should I stick with 1.5?
>>
>>101137773
Mangled, SD3 medium
>>
File: bedroom2.jpg (48 KB, 1152x832)
48 KB
48 KB JPG
>>
File: tqks-KRDSRFEzhmFSJFHW.png (1.07 MB, 1152x768)
1.07 MB
1.07 MB PNG
Safest model
>>
>>101137859
>human ears
>>
>>101137879
>>
File: CuIBKRJgc1bCHAx6dofBv.png (1.23 MB, 1024x1024)
1.23 MB
1.23 MB PNG
>>101137868
>>
File: sd3bo_00002_.png (2.49 MB, 1024x1728)
2.49 MB
2.49 MB PNG
>>
>>101137879
Give me five minutes if it bothers that much.
>>101137888
Oh, I thought I'd fix them in photoshop lmao.
>>
File: 00009-2423508070.png (2.81 MB, 1280x1920)
2.81 MB
2.81 MB PNG
>>101137899
Took me two seconds in iopaint
>>
File: bedroom2_fix.jpg (131 KB, 1152x832)
131 KB
131 KB JPG
>>101137888
Here's photoshop fix, now you can compare which one is better.
>>
>>101137843
Do both please. and 3 if you can.

You can make SDXL look realistic but theres just more available for 1.5.
I'm a sdxl user since release. SD3 does the best eyes, lips and hair.
>>
>>101137913
I like traditional rub and tuck service!
>>
File: sd3bo_00004_.png (2.77 MB, 1024x1728)
2.77 MB
2.77 MB PNG
>>
>>101137790
Not again!
>>
File: bedroom3.jpg (47 KB, 1152x832)
47 KB
47 KB JPG
This one needs an upscaling with sd15.
>>
File: snuggle-attack-racoon.gif (3.81 MB, 640x480)
3.81 MB
3.81 MB GIF
how can you bypass ai image upload p0rn detection?
>>
File: 07108-270739029.png (1.17 MB, 1344x768)
1.17 MB
1.17 MB PNG
>>
>>101137980
noise edits aren't working
>>
File: 00026-TFT_1249566.png (752 KB, 768x1280)
752 KB
752 KB PNG
>>
File: 1719258414738.jpg (102 KB, 768x1152)
102 KB
102 KB JPG
>>
File: 00081-TFT_1249565.png (654 KB, 768x1280)
654 KB
654 KB PNG
>>
File: 07114-3170533587.png (1.16 MB, 1344x768)
1.16 MB
1.16 MB PNG
>>
>>101137980
What are you uploading and to where?
>>
>>101137980
Can you explain what are you talking about?
>>
>Ani is crying
>>
>Ani is crying
>>
File: YEOsT6fTRECPWgJhZ6pKy.png (1.18 MB, 1024x1024)
1.18 MB
1.18 MB PNG
>>101137892
>>
File: sd3bo_00034_.png (3.04 MB, 1024x1728)
3.04 MB
3.04 MB PNG
>>
>>101138051
your typical onlyfans girl underwear some tits or ass, and the upload fails because of the lewdness, some works some not, the most obvious ones don't
>>
>>101138080
>>101138077
Jinx :]
>>
>>101138088
Overcooked and blurry as fuck
>>
File: ComfyUI_temp_inuzp_00052_.jpg (416 KB, 1200x1800)
416 KB
416 KB JPG
>>
>>101138093
I think your most typical is an issue between your ISP and Cloudflare.
I am posting from FI using fiber and I never see disappearing posts. I do not think this happens if I switch to 4g altough I do not post to 4chan that much using it.
>>
>>101138075
I believe the Anon isn't thinking Safety comes first. it's our upmost priority.
>>
>>101138110
>>101138093
Sorry, from Finland.
You should check out your browser and its cache etc.
>>
>>101138110
it's not here, the page has an image upload function that audits uploads
>>
>>101138130
What page?
>>
>>101138126
thanks btw

>>101138140
dalle or similar sites like that, it's not really as good but if I could get throught that I could get to the one I want, so any guideline would be appreciated
>>
>>101138153
I don't think you can.
God forbit if you take a photo of your own kid swimming in a lake Google's and Apple's AI will flag you as a pedophile.
>>
Next thread
>>101138197
>>101138197
>>101138197



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.