[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: 1703071746030150.png (1.07 MB, 1072x1072)
1.07 MB
1.07 MB PNG
Previous /sdg/ thread : >>101834483

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

>Local install
Automatic1111: https://github.com/automatic1111/stable-diffusion-webui
ComfyUI: https://github.com/comfyanonymous/ComfyUI
SD.Next: https://github.com/vladmandic/automatic
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

>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://aitracker.art
https://openmodeldb.info

>Black Forest Labs: Flux
https://huggingface.co/black-forest-labs/FLUX.1-schnell
https://huggingface.co/spaces/black-forest-labs/FLUX.1-schnell
https://comfyanonymous.github.io/ComfyUI_examples/flux

>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
>>
File: FDG_News_000034_.jpg (712 KB, 1344x960)
712 KB
712 KB JPG
>mfw Resource news

08/11/2024

>Plush-for-ComfyUI: Mixture of Agents and other Agentic Systems Workflows
https://github.com/glibsonoran/Plush-for-ComfyUI

>ComfyUI-SEGAttention: Smoothed Energy Guidance Node
https://github.com/logtd/ComfyUI-SEGAttention

>SimpleTuner v0.9.8.1 - much-improved flux training
https://github.com/bghira/SimpleTuner/releases/tag/v0.9.8.1

>Forge - BitsandBytes Guidelines and Flux [NB4]
https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/981

>ComfyUI_bitsandbytes_NF4
https://github.com/comfyanonymous/ComfyUI_bitsandbytes_NF4

08/10/2024

>Forge 2.0.0
https://github.com/lllyasviel/stable-diffusion-webui-forge/commit/cfa5242

>X-Labs Flux LoRA Collection
https://huggingface.co/XLabs-AI/flux-lora-collection
https://huggingface.co/Kijai/flux-loras-comfyui/tree/main/xlabs

>SwarmUI 0.9.2 Beta: Flux, SAM2, UI Improvements
https://github.com/mcmonkeyprojects/SwarmUI/releases/tag/0.9.2-Beta

>ComfyUI v0.0.5 Flux improvements and Hunyuan DiT controlnets
https://github.com/comfyanonymous/ComfyUI/releases/tag/v0.0.5

08/09/2024

>ComfyUI-Veevee: Video2Video Framework for ComfyUI
https://github.com/logtd/ComfyUI-Veevee

>BIGstockimage2M: ~2M stock images and captions @ ~0.4MP
https://huggingface.co/datasets/bigdata-pw/BIGstockimage2M

>Pitchfork Dataset: ~14 thousand album reviews and scores
https://huggingface.co/datasets/bigdata-pw/Pitchfork

>ComfyUI-LLaVA-OneVision
https://github.com/kijai/ComfyUI-LLaVA-OneVision

>AMD asks developer to take down ZLUDA, dev vows to rebuild
https://www.tomshardware.com/pc-components/gpus/amd-asks-developer-to-take-down-open-source-zluda-dev-vows-to-rebuild-his-project

>Freeway Animation HunYuan Demo: First large-scale Hunyuan-DiT fine-tuning
https://huggingface.co/Laxhar/Freeway_Animation_HunYuan_Demo

>DreamBooth training example for FLUX.1 [dev]
https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/README_flux.md
>>
>mfw Research news

08/11/2024

>JambaTalk: Speech-Driven 3D Talking Head Generation Based on Hybrid Transformer-Mamba Language Model
https://arxiv.org/abs/2408.01627

>SAT3D: Image-driven Semantic Attribute Transfer in 3D
https://arxiv.org/abs/2408.01664

>GLDiTalker: Speech-Driven 3D Facial Animation with Graph Latent Diffusion Transformer
https://arxiv.org/abs/2408.01826

>Latent-INR: A Flexible Framework for Implicit Representations of Videos with Discriminative Semantics
https://arxiv.org/abs/2408.02672

>Working with Color: How Color Quantization Can Aid Researchers of Problematic Information
https://arxiv.org/abs/2408.03424

>Decoding Visual Sentiment of Political Imagery
https://arxiv.org/abs/2408.04103

>Rethinking Visual Content Refinement in Low-Shot CLIP Adaptation
https://arxiv.org/abs/2407.14117

08/10/2024

>iControl3D: An Interactive System for Controllable 3D Scene Generation
https://arxiv.org/abs/2408.01678

>A General Ambiguity Model for Binary Edge Images with Edge Tracing and its Implementation
https://arxiv.org/abs/2408.01712

>TS-SAM: Fine-Tuning Segment-Anything Model for Downstream Tasks
https://arxiv.org/abs/2408.01835

>DeMansia: Mamba Never Forgets Any Tokens
https://arxiv.org/abs/2408.01986

>On Using Quasirandom Sequences in Machine Learning for Model Weight Initialization
https://arxiv.org/abs/2408.02654

>Decorrelating Structure via Adapters Makes Ensemble Learning Practical for Semi-supervised Learning
https://arxiv.org/abs/2408.04150

>LogogramNLP: Comparing Visual and Textual Representations of Ancient Logographic Writing Systems for NLP
https://arxiv.org/abs/2408.04628

08/09/2024

>Puppet-Master: Scaling Interactive Video Generation as a Motion Prior for Part-Level Dynamics
https://vgg-puppetmaster.github.io/

>Enhanced Prototypical Part Network (EPPNet) For Explainable Image Classification Via Prototypes
https://arxiv.org/abs/2408.04606

>Img-Diff: Contrastive Data Synthesis for Multimodal LLMs
https://arxiv.org/abs/2408.04594
>>
>loaded img
>>
File: delux_nf_00044_.png (2.25 MB, 1152x1344)
2.25 MB
2.25 MB PNG
>>101843516
ty wonderful baker
>>
File: 1723422428871.jpg (247 KB, 1024x1024)
247 KB
247 KB JPG
Neat candle
>>
File: artists.jpg (34 KB, 512x600)
34 KB
34 KB JPG
>>
File: ComfyUI_06060_.png (940 KB, 1752x1200)
940 KB
940 KB PNG
>>101843561
>first result
>>
>>101843599
The hands where better before.
>>
>>101843561
I fucking hate this character.
>>
>>101843516
How do I train a lora with prodigy? It always starts low and jumps too high for me. Is this normal?
>>
>>
>>101843635
yup. I believe I need to mess with the noise.
>>
>>101843657
whats the problem lol
>>
File: delux_nf_00051_.png (2.18 MB, 1344x1152)
2.18 MB
2.18 MB PNG
>>
>>
>>101843671
>prodigy
>dAdapt
it's working as advertised. you might need to consider other optimisers & schedules if this is a problem for you
>>
>>101843784
1 image, 100 steps. Almost every time it turns into this mess at 50-60 steps in prodigy right after the ramp up.
>>
File: 1723422480556.jpg (272 KB, 1024x1024)
272 KB
272 KB JPG
>>
>>
File: PW_81745_.png (1.2 MB, 1440x1024)
1.2 MB
1.2 MB PNG
Good evening, anons! I hope everyone is doing well :]
>>
File: 1723422437804.jpg (261 KB, 1024x1024)
261 KB
261 KB JPG
>>
File: PW_81757_.png (1.55 MB, 1440x1024)
1.55 MB
1.55 MB PNG
>>
>>
>>
File: 00023-4165916925.jpg (272 KB, 973x1252)
272 KB
272 KB JPG
>>101843939
Howdy
>>
File: delux_nf_00002_.png (2.29 MB, 1344x1152)
2.29 MB
2.29 MB PNG
>>101843939
hi there. short day?
I've been trying to make elementals but they've mostly just turned out as succubi with different tints
>>
File: 00134-3777061566.jpg (580 KB, 1024x1536)
580 KB
580 KB JPG
oldie
>>101843939
alright. it is thundering outside, kinda scared
>>
File: tmptm_93xbv.png (1.95 MB, 1024x1024)
1.95 MB
1.95 MB PNG
>>
File: PW_81734_.png (1.17 MB, 1440x1024)
1.17 MB
1.17 MB PNG
>>101844183
Hey there, anon! Great to see you :]
Cute gen!
>>101844207
Hey there, Debo! Yeah haha it was so boring, nobody came in so I just left early!
Hahaha it does look like a succubus but still really nice! :]
>>101844214
Oh no, I hope it's not too bad out! I would kill for anything besides just being hot out all the time LOL
>>
File: 1723422494593.jpg (228 KB, 1024x1024)
228 KB
228 KB JPG
>>
I just got Flux running. How do I prompt for a speech bubble or title?
>>
You lost
>>
File: 00000-1902759596.jpg (93 KB, 512x768)
93 KB
93 KB JPG
>>101844408
another oldie
godnat, venner og fjender
tomorrow, tomorrow is another day to improve, gen-wise, to innovate, to inspire, to belong in a positive, supportive, community of friendly a.i. (artificial intelligence) enthusiasts.
>>
File: Flux_00373_.png (1.41 MB, 1536x640)
1.41 MB
1.41 MB PNG
>>101844399
Usually just prompting something like
>Girl saying "I like cats"
usually works for me. But you can probably specify things like color/location of text if you want.
>Text at top of image that says "XYZ"
should work fine. You can also do things like
>Sign on top left of isle that says "Pickles"
>>
File: delux_nf_00010_.png (2.47 MB, 1344x1152)
2.47 MB
2.47 MB PNG
>>101844221
jokes on you, I was never buying a 5090 in the first place

>>101844262
do you actually get to wear the button up chef thingie?

>>101844388
https://suno.com/song/f5fb7035-6e82-487d-b2d1-629ab58df5f4

>>101844399
I usually just do "speech bubble with text "TEXT"" and it works well enough
>>
File: 1723422502434.jpg (306 KB, 1024x1024)
306 KB
306 KB JPG
>>
>>101844466
>>101844459
Thanks, I'll try both.
>>
>>101844399
> the words "TEXT" written {insert format/location info}
> the phrase "TEXT" ...
>>
File: PW_81742_.png (1.34 MB, 1440x1024)
1.34 MB
1.34 MB PNG
>>101844466
Yup! Though I choose not to wear it during the summer hahaha! It's already way too hot without it
I have a couple in black :]
>>
File: 00008-2931984805.png (464 KB, 768x512)
464 KB
464 KB PNG
>>101844519
For some reason it changed loicense to loliense, but pretty good.
>>
File: image.png (251 KB, 512x600)
251 KB
251 KB PNG
>>101843596
>>
File: 1723422509235.jpg (241 KB, 1024x1024)
241 KB
241 KB JPG
>>
File: ComfyUI_06066_.png (951 KB, 1752x1200)
951 KB
951 KB PNG
>>101843599
>>
File: ComfyUI_00203_.png (1.35 MB, 1152x896)
1.35 MB
1.35 MB PNG
>>
File: PW_81760_.png (1.92 MB, 1440x1024)
1.92 MB
1.92 MB PNG
>>101844466
Oh!! I just listened to your song! I'm making game music too hahaha! I hella like the style of yours!
https://suno.com/song/96703218-3a62-4db9-a4f1-ecc119903b0a
I think I posted this one last night but idk if you heard it haha
>>
File: delux_ak_00003_.png (3.89 MB, 1152x2016)
3.89 MB
3.89 MB PNG
>>101844659
this is a cool aesthetic. more of these?

>>101844664
thanks! I was listening to some of these yesterday but they were a little too bright and sugary for me, lol
>>
File: 00035-1691746925.png (1.01 MB, 896x1152)
1.01 MB
1.01 MB PNG
>>
File: PW_81756_.png (1.09 MB, 1440x1024)
1.09 MB
1.09 MB PNG
>>101844729
Hahaha fair enough! How about this one? I just made it :]
https://suno.com/song/f4220528-64df-4102-a32c-c70e7c44df8d
>>
File: delux_ak_00005_.png (3.03 MB, 2016x1152)
3.03 MB
3.03 MB PNG
>>101844767
epic boss music! I could practically hear the boss issues some melodramatic threat at the beginngin xD
>>
File: ComfyUI_00208_.jpg (738 KB, 2304x1792)
738 KB
738 KB JPG
>>101844729
>>
gn all
>>
File: PW_81776_.png (1.21 MB, 1440x1024)
1.21 MB
1.21 MB PNG
>>101844798
Hahaha I was totally thinking the same!! I'm glad you like it!
I've been saving all of the good ones to a playlist so I can find em easier for when I eventually add music to my game
>>
File: delux_ak_00008_.png (2.49 MB, 2016x1152)
2.49 MB
2.49 MB PNG
>>101844827
sick. that prompt works really well

>>101844838
gn

>>101844879
did you make much progress with the text-to-3d stuff? did the outputs end up being useful or nah?
>>
File: 00038-2188661947.png (1.3 MB, 896x1152)
1.3 MB
1.3 MB PNG
>>
File: 1723422589999.jpg (174 KB, 1024x1024)
174 KB
174 KB JPG
>>101844838
Gn anon
>>
File: ComfyUI_06069_.png (1.18 MB, 1752x1200)
1.18 MB
1.18 MB PNG
>>101844612
>>
File: PW_81800_.png (1.54 MB, 1440x1024)
1.54 MB
1.54 MB PNG
https://suno.com/song/814159c9-ada3-4248-83e1-d653237ae94f
>>101844897
LOOL nice gen
I didn't really use it for anything yet haha
Though i'm sure it's gonna work out really well! I'm just learning Unreal and getting basic mechanics down right now
>>
File: ComfyUI_00212_.png (3.2 MB, 2048x2048)
3.2 MB
3.2 MB PNG
>>
File: 00039-2317572949.png (1.66 MB, 1152x896)
1.66 MB
1.66 MB PNG
>>
File: PW_81829_.png (942 KB, 1024x1024)
942 KB
942 KB PNG
>>
File: ComfyUI_00217_.png (1.52 MB, 896x1152)
1.52 MB
1.52 MB PNG
>>
File: delux_ak_00010_.png (2.96 MB, 2016x1152)
2.96 MB
2.96 MB PNG
>>101844976
damn this gen is pretty crazy. some of the characters got a bit shuffled up but it got quite close. the CRT aesthetic is extra cool. who was the rightmost character supposed to be?

>>101845092
>debo's anti-hero arc
>>
File: PW_81825_.png (1.11 MB, 1024x1024)
1.11 MB
1.11 MB PNG
>>101845126
Thanks so much! :D
Yeah haha it does that a lot
Uhmm honestly it kinda just threw a random 5th person in there LOL
>debo's anti-hero arc
LOOL I didn't know what class to make you so I just picked thief
>>
>**** schizo arc
>>
File: 00042-3346202453.png (1.78 MB, 1152x896)
1.78 MB
1.78 MB PNG
>>101845092
Wish I still had my Witch Party CRT TV
>>
File: PW_81851_.png (2.12 MB, 1440x1440)
2.12 MB
2.12 MB PNG
>>101845226
Same! I just have this lame Inexlla Piry :[
>>
File: 00043-3558318616.png (1.7 MB, 1152x896)
1.7 MB
1.7 MB PNG
>>101845257
Dang, it's kind of smol
>>
File: PW_81860_.png (2.08 MB, 1440x1440)
2.08 MB
2.08 MB PNG
>>101845306
Very hahaha!
That looks really cool! I like the style you're doing
>>
File: ComfyUI_01070_.png (1.92 MB, 1600x960)
1.92 MB
1.92 MB PNG
good morning puter friends
>>
File: PW_81878_.png (2.27 MB, 1440x1440)
2.27 MB
2.27 MB PNG
https://suno.com/song/2a15c39e-e8b7-469f-88c6-dbb79476a170
>>101845417
Good morning, anon! I hope you've slept well :]
>>
File: 00003-1320673784.png (1.91 MB, 1152x896)
1.91 MB
1.91 MB PNG
>>101845341
Thanks. The RPG gens look good.
>>101845417
Hello
>>
File: delux_ak_00012_.png (3.18 MB, 1536x1152)
3.18 MB
3.18 MB PNG
>>101845417
hello

>>101845463
2v1 not fair
>>
File: PW_81874_.png (2.14 MB, 1440x1440)
2.14 MB
2.14 MB PNG
>>101845476
You're like twice the size and have cool battle music hahaha!
>>
>>101843516
Any tips for different art styles? I've tried 'realistic illustration', 'anime', and simple real life generation. I'm looking for new animation styles to mess around with.
>>
File: PW_81887_.png (1.87 MB, 1440x1440)
1.87 MB
1.87 MB PNG
>>101845548
I like doing 1930s~1950s cartoon styles sometimes haha and right now i'm just doing old rpg game stuff
>>
File: 00004-2574879046.png (1.93 MB, 1152x896)
1.93 MB
1.93 MB PNG
>>101845548
what are you using? That anime Flux lora is actually pretty decent
>>
>>101845618
Shalom
>>
File: 1672904272473739.png (1.84 MB, 1088x1408)
1.84 MB
1.84 MB PNG
>>101845618
Omg this actually just reminded me of a super old gen Fran made!
>>101845699
Um hello anon hahaha
>>
File: horse1.png (1.04 MB, 1024x1024)
1.04 MB
1.04 MB PNG
Is Flux lacking something? It does nice hands and a coherent image but its kind of boring. I felt like SDXL was more creative, even when it made abominations
>>
File: 00005-1093593939.png (1.76 MB, 1152x896)
1.76 MB
1.76 MB PNG
>>
>>101845746
I don't know that base was. Then again, I haven't used XL base since it first came out.
>>
File: delux_ak_00013_.png (2.57 MB, 1536x1152)
2.57 MB
2.57 MB PNG
>>101845746
I think there's a brewing consensus that flux is generally less creative and has some strong aesthetic biases that can leave gens feeling similar or bland
>>
Last one from me good night anons
>>
File: delux_ak_00014_.png (2.77 MB, 1536x1152)
2.77 MB
2.77 MB PNG
>>101845827
gn. take a plate for the road
>>
>>101845634
I am using nf4. I don't mind the anime style. I just want to experiment with other art types, some of which I may not know the name of.
>>
File: horsetay.png (1.58 MB, 1018x1018)
1.58 MB
1.58 MB PNG
>>101845746
>>101845746
this is after SDXL treatment
i like making surreal shit like undead taylor swift riding a 5 legged horse
>>
File: 00008-1420367602.png (1.71 MB, 1152x896)
1.71 MB
1.71 MB PNG
>>101845859
>It's not a flaming horse
>>
File: PW_81904_.png (1.56 MB, 1440x768)
1.56 MB
1.56 MB PNG
>>101845827
Good night, Quokkanon! Sleep well :]
>>
File: 00208-174680235.png (1.77 MB, 1152x896)
1.77 MB
1.77 MB PNG
>>
Suno refuses to prompt a song with "Sayonara" kek. Has censorship gone too far?
>>
File: tayhorsefail.png (1.88 MB, 1017x1017)
1.88 MB
1.88 MB PNG
>>101845879
picrel is my failed attempt
if anyone can make a proper flaming horse i will be impressed
>>
>>101845956
Is he alright?
>>
>>101846015
that's why I want a suno/udio tier model on local, I'm tired on walking on eggshells when trying out cucked APIs
>>
File: PW_81913_.png (1.33 MB, 1440x768)
1.33 MB
1.33 MB PNG
>>101845956
This looks amazing!
>>101846015
Wow really? LOL
That's so weird, have you tried writing it in Japanese?
>>
File: 00011-1224941860.png (1.59 MB, 1152x896)
1.59 MB
1.59 MB PNG
>>101846017
Nice attempt though.
>>
>>101846034
He lives.
>>101846057
Thanks
>>
>>101846057
Oh yeah I have tons! Too many in fact.
https://suno.com/song/7defced2-5578-461d-a311-1aec3d309636
https://suno.com/song/68c61b21-aa6b-4031-a1df-328d4392b0fc
>>
File: delux_ak_00018_.png (3.02 MB, 1536x1152)
3.02 MB
3.02 MB PNG
>>
File: 00402-2234203.png (1.56 MB, 1152x896)
1.56 MB
1.56 MB PNG
>>
File: PW_81960_.png (1.91 MB, 1440x1024)
1.91 MB
1.91 MB PNG
>>101846177
Oh hey it's Mouse anon!! :D
Helloooooo!
Those sound great! My first song on Suno was actually in Japanese haha!
https://suno.com/song/9856bf3c-7bd9-4945-badd-c384ff0173dc
>>101846017
This is the best attempt I have so far haha! I'm gonna keep giving it a try!
>>
>>101846236
cool, rompt?
>>
>>101843516
I've been trying to get my hands on some japanese stuff since it seems that they finally managed to feed enough anime images that it's finally generating the good stuff
>>
File: PW_81966_.png (2.15 MB, 1440x1024)
2.15 MB
2.15 MB PNG
Oh wait this one is a bit better!
>>
File: delux_il_00035_.png (3.42 MB, 1536x1152)
3.42 MB
3.42 MB PNG
i slep

>>101846177
cute

>>101846236
cool

>>101846256
>>101846265
cool (hot)

>>101846264
gl?
>>
>>101846258
https://files.catbox.moe/b1c2sp.png
You could probably distill the prompt down. Been a while since I've played around with image diffusion. I'm using words from old prompts to feel out flux.
>>
File: PW_81975_.png (2.21 MB, 1440x1024)
2.21 MB
2.21 MB PNG
>>101846290
Thanks!! :D
Night Debo! Sleep well!
>>
File: 00014-1285086854.png (1.85 MB, 1152x896)
1.85 MB
1.85 MB PNG
>>101846265
It handles multiple concepts better than I expected
>>
boring realism lora for flux dev is quite good
>Bathroom selfie from Donald Trump. Harsh fluorescent lighting.
>>
File: 00015-2280502627.png (1.85 MB, 1152x896)
1.85 MB
1.85 MB PNG
>>101846358
The xlabs one? It really helped quality more than I thought it would
>>
>>101846391
no the xlabs one sucks shit

this one:
https://civitai.com/models/639937?modelVersionId=715729
>>
>>101846256
Wow nice I like this a lot. Also I misread, Japanese prompt in this one works with sayonara dunno why it doesn't work with English, super weird indeed.
https://suno.com/song/aaf2a1d3-55c2-4698-ad74-60ca604a4eb0
>>
File: file.jpg (279 KB, 1536x1536)
279 KB
279 KB JPG
>>
File: F_00070_.png (993 KB, 968x1088)
993 KB
993 KB PNG
>>101846399
I haven't tried that one. Downloading it now.
This was without the xlabs one
>>
File: PW_81981_.png (2.02 MB, 1440x1024)
2.02 MB
2.02 MB PNG
>>101846347
I had to reeeeeally tell it what to do LOL
>riding on the back of a horse engulfed in flames, a horse made out of fire, hooves leave burning trail
I used FluxDev, Schnell wasn't burning the horse haha
>>101846417
Thanks! :D
It's about not wanting to wake up and stay in bed hahaha!!
Huh, that is weird
I can't see literally any reason why that would be censored
This sounds great! :D
>>
File: 00001-3410670811.png (1.18 MB, 896x1152)
1.18 MB
1.18 MB PNG
>>101846462
This was with. I thought it gave it a noticeable quality boost.
>>
File: BMP_FLUX_06105_.png (952 KB, 1024x1024)
952 KB
952 KB PNG
>>101846470
Yeah it reminded me of this
https://www.youtube.com/watch?v=Yv6shy_9KVM
I watch too much youtube, sometimes the simplistic songs are the best if you wanna rock out
>>
File: BMP_FLUX_05512_.png (762 KB, 1024x1024)
762 KB
762 KB PNG
>>
File: PW_80614_.png (1.15 MB, 1440x512)
1.15 MB
1.15 MB PNG
>>101846561
LOOOOL this is cute! I love it hahaha
>>
File: Capture.jpg (328 KB, 2875x1521)
328 KB
328 KB JPG
Holy fuck, I found a node "SEGAttention" that gives good output for CFG 6 without the need of DynamicThresholding
https://github.com/logtd/ComfyUI-SEGAttention
>>
File: PW_80625_.png (1.73 MB, 1536x768)
1.73 MB
1.73 MB PNG
Good night anons! I hope to see you all tomorrow :]
Today has been fun
>>
File: Capture.jpg (327 KB, 3001x1342)
327 KB
327 KB JPG
>>101846688
Did some test between DynamicThreshold and SEGAttention
>CFG 6
>Guidance Neg 10
>Prompt: An illustration inspired by the works of Jean-Baptiste-Siméon Chardin. The scene depicts a 18th-century classroom with Hatsune Miku as a teacher at the center, guiding a group of attentive young students. Miku is dressed in her classic outfit altered to fit to the setting. The classroom is filled with wooden desks, chalkboards, and books, all rendered with Chardin's characteristic focus on realistic textures and warm, muted colors. The words "How to prooompt" are clearly written on the chalkboard. There is a speech bubble coming from Miku's mouth with the text "Don't be a promptlet". The lighting is soft and natural, creating an atmosphere of calm and scholarly dedication.
https://imgsli.com/Mjg2Njkz
DynamicThreshold seems to win but SEGAattention has the default values there so it can be improved I guess, more testing needs to be done
>>
File: 01266-741534496.png (839 KB, 756x948)
839 KB
839 KB PNG
Ow my fucking ears
https://suno.com/song/3dac4420-7e17-4d63-960a-d94b354f16f4
>>
>>101846813
good night peep
>>
>>101846838
https://suno.com/song/1df84b54-9bda-4b3d-997d-ae59a61eb1b9
>>
>>101846813
Take it easy dude!
>>
>>101843596
Bold. Powerful. Not my problem.
>>
File: file.jpg (188 KB, 1024x1024)
188 KB
188 KB JPG
gm
103m flickr
>>
>>101846816
false flag, SEGAattention gives the same output whether we're on cfg = 6 or cfg = 1, but on cfg = 1 it can do negative prompt though, and it's not 3x slower like on PerpNegAdaptiveGuider, it's the regular speed you get on cfg > 1
https://imgsli.com/Mjg2Njk4
>>
>ded schizo containment general
>>
>>101844976
>4/4 in prompt
>song is in 5/4
>>
File: ComfyUI_01092_.png (2.61 MB, 1600x960)
2.61 MB
2.61 MB PNG
>>
File: warp_131.jpg (199 KB, 1312x1312)
199 KB
199 KB JPG
>>
File: 1723453804959972.jpg (303 KB, 2711x1321)
303 KB
303 KB JPG
Oh shit, Euler CFG++ actually works on flux, you just need this node
https://github.com/pamparamm/ComfyUI-ppm
https://imgsli.com/Mjg2NzE2
>>
>>101847528
>boomer prompting
>no artistic idea whatsoever
Yeah this is why normies love their toys.
>>
>>101847528
When you need to larp a failed novelist, it's a sign this technology has gone off the rails.
>>
>>101847851
you can get way better results than that, was just the basic CFG 1 test there, here's how far you can go: https://imgsli.com/Mjg2Njkz
>>
>>101847880
Maybe you didn't get the subtle hint.
>>
>>101848101
I don't really give a fuck about your hint anon
>>
>>101848117
but you still replied. seems like schizoanon wins this time
>>
>>101848140
Can you shut the fuck up for a single day?
>>
>/ldg/ won
>schizoanon won
/sdg/ sisters...
>>
schizo containment thread
>>
File: ComfyUI_01010_.png (1.72 MB, 1600x896)
1.72 MB
1.72 MB PNG
>>
I've read here that adding multiple LORAs tend to result in overcooked generations.
If you add your LORAs at 0.7 weight but give the trigger words 1.2 weight, will the result be better?
>>
>schizophrenic
>>
>
>>
>>101848827
It's always better to avoid any loras because they will practically lobotomize the model's understanding.
Sure most slop generators don't even need a lora for their 1girl, photoreal, big boobs -gens anyway.
If you are so concerned about the question why don't you do some tests and report back?
>>
>>101849045
>It's always better to avoid any loras because they will practically lobotomize the model's understanding
I used loras too often and now my model is legally retarded, I prompt 1girl and it gives me downies. It's over.
>>
File: 000000_16243_.png (2.56 MB, 1075x1434)
2.56 MB
2.56 MB PNG
G'mornin Anons, have a great day!
>>
>>101849436
kys
>>
>>101849511
good morning

>>101849526
why, what did I do?
>>
>>101848827
test it
first of all you need to define "overcooked", think about the criteria you'll use to rate the results
collect a reasonable sample of loras taking time and compute into account
choose a sample of prompts and decide the other base generation parameters that you'll keep the same
decide the ranges of prompt weights/lora weights and the range of loras to combine, taking into account the resulting number of combinations and how that affects the required time/compute

tl;dr whatever you wanted to hear, yes/no/maybe
>>
>mfw debo
>>
ran took everything from me
>>
Comfyanon's discord wants to doxx everyone because they are no-skilled autists. There is a reason why certain people have been posting literally the same subject and image for two years by now.
>>
>>101849675
>>101849752
>>101849799
see >>101848232
>>
>thread's still alive

>>101844838
>>
>>
>>
>>101850040
we just fake activity here so debo stays in his containment thread
>>
>>101850167
my activity is genuine
>>
>>101850209
Dont worry! All newfags learn at some point
>>
We won
>>
>mfw
>>
>>
File: IMG_20240812_162443.jpg (668 KB, 1007x1433)
668 KB
668 KB JPG
>>
>>101850700
>>101850683
thank you for your service
>>
>>
>>101850781
is it wrong to post gens itt?
>>
>>101850999
Absolutely not.
>>
>>101850999
Seems like you're not part of the group. Dont worry! Just keep posting here anon
>>
>>
I haven't tried upscaling at all and there are a billion models now, what would be a good place to start for upscaling low-resolution realistic textures from an old game? Grass, flowers, clouds, etc. 4x-Nomos8kDAT? StableSR?
>>
File: file.jpg (413 KB, 1792x1024)
413 KB
413 KB JPG
when you come out your shit is gone
>>
>>101851119
Start with Ersgan and get your pipeline ready first. Then you know what you need after playing around a bit. It's never a linear straightforward process. This is also texture specific.
>>
>>101851129
Did they nuke Big Data on Huggingface?
>>
File: file.jpg (403 KB, 1792x1024)
403 KB
403 KB JPG
BIG things happening soon:tm:
keep it locked, stay noided
>>101851229
i dont think so, still active for me and visible from private browsing window. why, can't you access it?
>>
>>
>>101851263
Why not post in the other thread?
Debo and ani used to insult you
>>
File: ComfyUI_01103_.png (2.18 MB, 1600x896)
2.18 MB
2.18 MB PNG
>>
>>101851292
:shrug: everyone did, i don't let things like that get to me any more. the other thread is more active desu
>>
>>101851292
He works for me and stays in this thread.
>>
File: ComfyUI_Flux_7115.jpg (201 KB, 1024x1024)
201 KB
201 KB JPG
>>
What am I doing wrong here? Why does it look so awful and out of focus with poor detail. I'm new to pony on A1111 using ponyRealismXL.
>>
>>101851614
>Resolution: 1280x768 | 768x1280
These apparently are the buckets that model was trained at, might have better clarity with those dims
>>
>>101851676
>>101851614
SDXL standard resolutions will suffice and there's some leeway always as long as it's not retarded.
It's your VAE perhaps, are you using sdxl vae? God knows what they have embedded into the checkpoint itself.
>>
>>
I am running Debian (Linux) with an AMD GPU.
Is it possible to install Stable Diffusion easily and successfully?
>>
>>101852137
possible? yes. easily? probably not.
>>
File: 323.jpg (47 KB, 600x722)
47 KB
47 KB JPG
i've had enough, how do i treat proooompting addiction, books on the subjects? or straight to therapy
>>
>>
>>
File: uyytuyu.jpg (3.68 MB, 2688x8448)
3.68 MB
3.68 MB JPG
>>101852272
might be the wrong crowd to ask
>>
>when you settle in court for $122 million
>wake up from dream shortly afterwards
I'm ngmi, gm tho
>>
What prompts or resources should I use on CivitAI to achieve link related?

https://img.booru.org/blacked//images/37/62f4254bb3af318b0e4e3c60e38286a94708117d.png
>>
File: delux_nf_00031_.png (582 KB, 1344x768)
582 KB
582 KB PNG
>>101852272
the secret is to become a functioning addict. people will think you're cool

>>101852306
gm
>>
File: ComfyUI_01116_.png (1.92 MB, 896x1344)
1.92 MB
1.92 MB PNG
>>
>>
>mfw debo
>>
>>101852700
where?
>>
>>
>>
>>101852959
big
>>
>>101852972
>>
>>
File: delux_nf_00030_.png (1.25 MB, 1344x768)
1.25 MB
1.25 MB PNG
>>
File: 00133-3898414299.jpg (86 KB, 512x768)
86 KB
86 KB JPG
>>
File: ComfyUI_00237_.png (1.96 MB, 1280x1280)
1.96 MB
1.96 MB PNG
>>
>>
File: ComfyUI_00239_.png (2.24 MB, 1408x1408)
2.24 MB
2.24 MB PNG
>>
File: delux_nf_00028_.png (1.92 MB, 1344x768)
1.92 MB
1.92 MB PNG
>>
>>101853124
For a moment i thought thats a miniani
>>
File: delux_nf_00026_.png (2.02 MB, 1344x768)
2.02 MB
2.02 MB PNG
>>101853472
lol, true, that gen did kinda nab his aesthetic. I think he always gens squares tho
we've been drastically short on animations lately :(
>>
>>101853537
Yeah he always did 380x380
I wonder what happened to him
>>
File: 00314-1502799687.jpg (186 KB, 512x768)
186 KB
186 KB JPG
>>
File: file.jpg (222 KB, 1024x1024)
222 KB
222 KB JPG
n-grams
>>
File: delux_nf_00025_.png (1.83 MB, 1344x768)
1.83 MB
1.83 MB PNG
>>101853592
its not unusual for him to disappear for weeks. he gets in the mood to gen a bunch and then we don't see him again for a while

>>101853628
very on-brand. have to remember to show this to pw
>>
File: lowskill.png (58 KB, 200x187)
58 KB
58 KB PNG
>>101853537
lol
>>
>lets talk about the heckin cute and valid avatarfags again
>>
Morning anons
>>
>>101853719
talk about me first
>>
File: delux_nf_00023_.png (1.72 MB, 1344x768)
1.72 MB
1.72 MB PNG
>>101853722
gm. did you ever try getting nf4 running on your machine? its surprisingly light-weight
>>
>>101853761
>Looks decent
>piss yellow filter
I'll give you a pass
>>
>>
File: delux_nf_00022_.png (1.91 MB, 1344x768)
1.91 MB
1.91 MB PNG
>>101853776
I couldn't figure out where the aggressive tinting was coming from in this series. I was playing around with cfg/guidance params a bunch but also the prompt was pretty overloaded too. the style was cool but, I agree, the coloring wasn't great
>>
It's 86f where I am and I don't have AC, hope it's cooler where you guys are. In the meantime, have a new thread...
>>101853740
>>101853740
>>101853740
>>
File: ComfyUI_01112_.png (1.57 MB, 1600x896)
1.57 MB
1.57 MB PNG
>>
File: ComfyUI_temp_saskn_00062_.png (1.61 MB, 1024x1024)
1.61 MB
1.61 MB PNG
>>
I love chubsters
>>
>>101853761
Haven't tried it but It's very likely my 1060 isn't compatible with NF4 but I did updated forge, I'll give it a try
>>
File: delux_nf_00021_.png (1.99 MB, 1344x768)
1.99 MB
1.99 MB PNG
>>101853823
>86f
relatable. go lounge in a coffee shop or something. sounds too hot to gen :( we have to fight climate change otherwise we won't be able to gen year round
>>
it's over
>>
>>101853938
don't be sad that it's over, be happy that theres freshly baked bread
>>101853740



[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.