[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: tmp.jpg (1.19 MB, 3264x3264)
1.19 MB
1.19 MB JPG
/ldg/ - Local Diffusion General

Discussion of free and open source text-to-image models

Previous /ldg/ bread : >>101433582

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

>Advanced UI
Automatic1111: https://github.com/automatic1111/stable-diffusion-webui
ComfyUI: https://github.com/comfyanonymous/ComfyUI
InvokeAI: https://github.com/invoke-ai/InvokeAI
SD.Next: https://github.com/vladmandic/automatic
SwarmUI: https://github.com/mcmonkeyprojects/SwarmUI

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

>Model Ranking
https://imgsys.org/rankings

>Models, LoRAs & training
https://civitai.com
https://huggingface.co
https://aitracker.art
https://github.com/Nerogar/OneTrainer
https://github.com/derrian-distro/LoRA_Easy_Training_Scripts

>Pixart Sigma & Hunyuan DIT
https://huggingface.co/spaces/PixArt-alpha/PixArt-Sigma
https://huggingface.co/spaces/Tencent-Hunyuan/HunyuanDiT
Nodes: https://github.com/city96/ComfyUI_ExtraModels

>Kolors
https://gokaygokay-kolors.hf.space
Nodes: https://github.com/kijai/ComfyUI-KwaiKolorsWrapper

>AuraFlow
https://fal.ai/models/fal-ai/aura-flow
https://huggingface.co/fal/AuraFlows

>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

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

>Related boards
>>>/h/hdg
>>>/e/edg
>>>/d/ddg
>>>/b/degen
>>>/vt/vtai
>>>/aco/sdg
>>>/trash/sdg
>>
One thing I've noticed that public AI image generators like Bing have an advantage over ones you install on your PC is that Bing for example, with access to internet, you can use terms and things that are much more niche and still get results, like D&D races such as firbolgs. Fooocus, however, can't do that.
>>
why /ldg/ /sdg/
>>
fake and gay
>>
File: PixArt_00001_.png (1.31 MB, 1280x768)
1.31 MB
1.31 MB PNG
PixArt workflow
https://pastebin.com/tnQKhTLT

SD3 PixArt workflow
https://pastebin.com/AJ9uJrmM
>>
Made the OP collage again. Nice.
>>
>>101468963
That's not how that works
>>
File: 0.jpg (378 KB, 1024x1024)
378 KB
378 KB JPG
>>
File: Sigma_05712_.png (1.35 MB, 1024x1024)
1.35 MB
1.35 MB PNG
>>
stale bred finally got tossed
>>
Here lies: LDG
>>
File: 0.jpg (262 KB, 1024x1024)
262 KB
262 KB JPG
>>
>>101473184
>Here lies: LDG
Did /ldg get crowdstriken?
>>
File: file.png (970 KB, 1280x768)
970 KB
970 KB PNG
>>
>>101468672
>/sdg/ - Stable Diffusion General
>/ldg/ - Local Diffusion General
kek
>>
>>101471377
so much airbrush
>>
What's the best easy setup for coloring in anime pictures with references? Anything on hugging face?
>>
>>101471854
thread be slow these days
>>
New model gimmie gimmie.
>>
>>101475630
well it's pixart, kolors and aura. plenty of people will tell it's "just pixart" so try each one for yourself and see what you like.
>>
>>101473309
lul gonna have to delete files from the system directory to save it
>>
ldg is dead long live ldg
>>
>>101476524
Baker just wasn't paying attention. Confusing as a schizo split though
>>
>>101476768
Baker also shit up sdg the other day. Bad look. True LDGers appreciate both threads.
>>
>>
>>
>>101468672
Anm i the only one who fidns all of the images in OP dull and tiresome?
>>
>>101477206
You're the only one who has said anything
>>
>>101477206
I agree but I don't bake
Depends on the baker desu
>>
File: file.png (175 KB, 352x176)
175 KB
175 KB PNG
You know, I realize I never tested inference. That would've been funny if I spent months and it didn't work.
>>
File: file.png (91 KB, 577x361)
91 KB
91 KB PNG
This is how I'll know the model is done training.
>>
>>101477601
kek also hype
>>
.
>>
>>101477601
>>101477641
>kek also hype
^
>>
Lost
>>
dont necrobump this dead thread
>>
where did it go wrong?
>>
File: Sigma_05795_.png (950 KB, 768x1280)
950 KB
950 KB PNG
1girl 1bump
>>
pointless thread without discussion or content, just let it die
>>
Would you rather discuss... I don't know... DXM, anon?
>>
>>101480697
the heck?
>>
File: Sigma_05867_.png (1.54 MB, 1024x1024)
1.54 MB
1.54 MB PNG
>>101480394
No new local models. I'm here to post images on an imageboard
>>
soon
>>
>>101480394
If you feel more at home with the underaged banned on SDG posting 1girls you can stay over there. It's not like /g/ is a fast moving board.
>>
if it were about speed, you'd be in /degen/ right now
>>
>>101473450
Model?
>>
>>101469734
>>101473184
>>101476524
>>101479698
>>101479808
>>101480394
gm anon
>>
>>101481025
it was some pixart checkpoint. can't launch any ui tho, training
>>
File: Sigma_05885_.png (1.25 MB, 768x1280)
1.25 MB
1.25 MB PNG
Bumpety
>>
File: Sigma_05887_.png (1.44 MB, 768x1280)
1.44 MB
1.44 MB PNG
It's a shame auraflow went the way that it did. Was rooting for cloneofsimu. At least bigma's cooking
>>
File: grid-0000.jpg (1.31 MB, 4000x2667)
1.31 MB
1.31 MB JPG
>>
File: Sigma_05892_.png (1.21 MB, 768x1280)
1.21 MB
1.21 MB PNG
>>
>>101481603
The worst part is it appears that he's simply continuing training as opposed to completely omitting synthetic images from the dataset and restarting. Many seem to like the prompt adherence however.
>>
>>101481835
He's a fucking retard because those images are basically burned into the model and might be impossible to wash out. That cat is going to haunt many images.
>>
File: grid-0001.jpg (1.24 MB, 4000x2667)
1.24 MB
1.24 MB JPG
>>
File: grid-0002.jpg (1.26 MB, 4000x2667)
1.26 MB
1.26 MB JPG
>>
>>101477492
What was the prompt for this one?
>>
>>101482339
graffiti on a brick wall
>>
File: 0.jpg (285 KB, 1024x1024)
285 KB
285 KB JPG
>>
Bigma ETA?
>>
File: file.png (141 KB, 256x256)
141 KB
141 KB PNG
>>101483112
I'm sure whatever the Pixart team is working on will come out within 6 months. If you're talking about 1.3B, who knows, I'm still trying to dial in the learning rate but the dataset is mostly done but I'm still captioning and throwing in 4archive/b
>>
Why is it so slow today? Is it because of people working overtime to fix the CrowdStrike fuckup?
>>
>>101483255
that or nothing really habbening since aura dropped
>>
>>101483214
>4archive/b
Nice. Anything funny/notable?
>>
>>101483326
no, but I filter out rekt and other weird shit
but there's still going to be some weird stuff in there including rare pepe threads
>>
>>101475886
>kolors
outdated architecture
>aura
undercooked and dataset fuck ups

our only hope is sigma...
>>
>>101477206
Why the stuff like >>101437575 wasnt included is beyond me.
>>
>>101481603
love foggy days
>>
File: pixart_00041_.png (1.06 MB, 1280x768)
1.06 MB
1.06 MB PNG
>>
do you use regularization images when training character loras? i've never tried and i've never really understood the point. and the regularization images i've downloaded have been extremely low quality which just seem like they would spoil the entire model
>>
>>101484908
not really but it doesn't hurt to just make our own dataset. just roll through the txt with the checkpoint you are using for training
>>
>>101484929
>just roll through the txt with the checkpoint you are using for training
oh yeah i remember generating a bunch of random images like that when i first started training. completely forgot about that. what do you mean iwht "roll through the txt" though?
>>
File: 0.jpg (93 KB, 1024x1024)
93 KB
93 KB JPG
>>
File: __00232_.png (837 KB, 896x1152)
837 KB
837 KB PNG
>>
>>101483112
>>101480787
>>
File: 000443-382695684.png (2.73 MB, 1072x1880)
2.73 MB
2.73 MB PNG
>>
>>101485256
Legible symbol
>>
File: __00260_.jpg (702 KB, 1792x2304)
702 KB
702 KB JPG
>>
File: __00262_.png (894 KB, 896x1152)
894 KB
894 KB PNG
:<)
>>
>>101471377
shes staring into my soul
>>
File: catbox_it5fv3.png (2.6 MB, 1152x1632)
2.6 MB
2.6 MB PNG
>>
File: the stare .png (248 KB, 540x570)
248 KB
248 KB PNG
>have several coworkers social medias
>download photos with their faces
>use face swapper to get a 90% accurate face
>generate soulless nude selfies until I coom
after work I just go home and jerk off to whoever I talked with the most that day. after several weeks it isnt really fun anymore because theres no thrill to fake nudes, yet I still cant help but do it
>>
>>101487452
>those gens? merely personal use; i would never distribute, officer.
>>
File: __00264_.png (1.04 MB, 896x1152)
1.04 MB
1.04 MB PNG
>>101487334
didnt even realize this was pony, i like it
>>101487452
infinite digital harem
>>
>>
>>
>>101468672
>>Models, LoRAs & training
add https://www.modelscope.cn/
>>
>>101481354
They're nervous
>>
File: catbox_n7esj1.png (1.72 MB, 1152x1632)
1.72 MB
1.72 MB PNG
>>101487740
Turns out that it's possible to train lora's in a way that it's legitimately hard to tell that something was made via base pony(even with 1girl, standing).
>>
>>
>>101481835
>completely omitting synthetic images from the dataset and restarting
that is absolutely retarded, please exit life
>>
er... unofficial pixart bigma and lumina 2 waiting room?
>>
>>101488812
link to that lora?
>>
>>101489127
https://mega.nz/file/03NhibrI#rVGNcMYHHAEIRPHt_bj49urOGu72_aKQU7P9Gu2PZG0
>>
>>101489212
ty!
>>
File: __00294_.png (1.07 MB, 896x1152)
1.07 MB
1.07 MB PNG
>>
File: __00301_.png (1.01 MB, 896x1152)
1.01 MB
1.01 MB PNG
>>
File: __00315_.jpg (890 KB, 1792x2304)
890 KB
890 KB JPG
>>101489212
using with -1 model and clip strength
>>
>>101490075
>>101489728
>>101489583
>>101487740
>>101486554
>>101486103
>>101485413
Anon this is amazing. Where can I download terrabytes of this stuff? Please dump everything you have. Or could you share a catbox?
>>
>>101490075
how i feel after accidentally inhaling water
>>
File: __00318_.png (1 MB, 896x1152)
1 MB
1 MB PNG
>>101490149
ty. i occasionally post to twitter but mainly here
https://files.catbox.moe/eemv16.png
>>
File: Sigma_05722_.png (1.37 MB, 1024x1024)
1.37 MB
1.37 MB PNG
>>101484328
Basically walking in the clouds on land

>>101487065
>shes staring into my soul
She means to take it
>>
>>101490317
Link to your twitter if you want to share it here?
>>
>>101490429
@meatcrowning
>>
>>101490149
this. Gen some background wallpapers? i like how dark and weird they are.
>>
>>101490500
NTA but dank posts
>>
I have a uniformed question that I suspect the answer is no too

Can I just buy a edge TPU and run it in tandem with my GPU for generation purposes? I feel like a GPU and a TPU should be able to work together since you can stack GPUs and have them work together, but on the other hand TPUs are pretty new and I find it highly unlikely that there would be any framework put down for them to be able to work in tandem with a gpu
>>
>>101490784
as far as i know those are not usefull for sd so far.
>>
Is there a SD discord for devs?
>>
>>
File: ComfyUI_0001.jpg (2.42 MB, 1664x2432)
2.42 MB
2.42 MB JPG
>>101491617
>>101491732
Wouldn't recommend it.
>>
>>101491617
Seems pretty generic for an unofficial discord. You'll probably find developers who use SD in discords for various projects. Pixart discord has lots of dev talk but rarely about SD
>>
File: Sigma_05385_.png (1.32 MB, 1280x768)
1.32 MB
1.32 MB PNG
>>
>>101491962
>>101491855
Just need help on how to ensure my postprocessing script runs after any other changes to the image are made.
>>
File: sdxl_18.jpg (285 KB, 1480x1328)
285 KB
285 KB JPG
What's the latest hype
>>
File: Sigma_05360_.png (1.38 MB, 1280x768)
1.38 MB
1.38 MB PNG
>>101492031
Have you tried throwing the script in https://claude.ai/ and asking questions?

>>101492052
Cascade got an upgrade https://github.com/2kpr/ComfyUI-UltraPixel
>>
Any good generic loras for a character who wears a screen over her face?
>>
>>101492107
Yeah, I do have other questions which AI isn't quite up to date to help me.
>>
>>101492198
Did you try just prompting? Using SDXL should work.
>>
>>101492538
Yes, it did nothing for what I'm looking for. Amazing how such a simple concept makes the AI spaz out unless I'm using a lora for Rina.
>>
File: 0.jpg (603 KB, 2048x1024)
603 KB
603 KB JPG
>>
>>101492031
Then take his advice >>101491962
>Pixart discord has lots of dev talk but rarely about SD
>>
>>
Good morning. I forgot to post so many "outstanding" gens. Can you post an outstanding gen with no control input?
>G0RWHP
>AX248M
>>
>>
so how do we download this sigma thing
>>
>solve 5 RP captchas to post with the wrong keyboard layout
>DM forces me to auth again
>DPRNHA
I'll post it I want.
>>
>>101487452

The very thought of doing that shit creeps me out. I stick to 3D, or drawn shit. 2.5 style is pushing it.
>>
File: comfyui_sdxl_00023_.png (1.38 MB, 1024x1024)
1.38 MB
1.38 MB PNG
>>
File: Sigma_06018_.png (1.47 MB, 1280x768)
1.47 MB
1.47 MB PNG
>>
File: Sigma_06028_.png (1.66 MB, 1280x768)
1.66 MB
1.66 MB PNG
>>
>>101492198
describe screen. TV screen, fencing mesh, lace face mask?

>>101492538
you should tell him to prompt for keyboards and hard edges next.
>>
File: ComfyUI_00767_.png (3.12 MB, 1280x1792)
3.12 MB
3.12 MB PNG
>>
File: ComfyUI_00967_.png (2.55 MB, 1280x1792)
2.55 MB
2.55 MB PNG
>>
File: ComfyUI_01161_.png (2.88 MB, 1280x1792)
2.88 MB
2.88 MB PNG
>>
File: comfyui_sdxl_00058_.png (1.29 MB, 1024x1024)
1.29 MB
1.29 MB PNG
>>
File: moooooo.jpg (561 KB, 1273x1142)
561 KB
561 KB JPG
Post your best cow gens
Hard mode: No big titty cow "girls"
>>
File: Sigma_06110_.png (1.59 MB, 1024x1024)
1.59 MB
1.59 MB PNG
>>101494286
Mooooon
>>
File: Sigma_06116_.png (1.62 MB, 1024x1024)
1.62 MB
1.62 MB PNG
>>101493203
Didn't see you there. Follow install for Sigma at https://github.com/city96/ComfyUI_ExtraModels
And grab a model from https://civitai.com/search/models?baseModel=PixArt%20E&sortBy=models_v9
>>
File: 0.jpg (107 KB, 1024x1024)
107 KB
107 KB JPG
>>
Civitai down for anyone else? Planned maintenance?
>>
>>101494439
Your workflow
>>101470318
>>
is it normal that my pc becomes unusable with freezes for a solid 15min+ when switching checkpoints in forgewebui?
>>
>>101468672
is ultrapixel good? can anon run it?
>>
>>101491962
>Pixart discord has lots of dev talk but rarely about SD
kek
>>
>>101495106
>ultrapixel good
Stable Cascade with upscale built in. Cascade ranks better than SD3, and worse than Sigma on imgsys.org
>>
>>101492107
neat
>>
Can someone with this shit set up generate me a picture of a muscular big chungus flexing together with anubis in the egyptian pantheon?
>>
>>101495999
Use bing for stuff like that
>>
File: 116775249263548999-SD.png (1.76 MB, 1232x960)
1.76 MB
1.76 MB PNG
major bumpage
>>
>>101495999
Pay me
>>101496198
Nice bump
>>
File: Sigma_06136_.png (1.61 MB, 1280x768)
1.61 MB
1.61 MB PNG
>>101495951
ty
>>
File: 116775249263549003-SD.png (2.8 MB, 1016x1648)
2.8 MB
2.8 MB PNG
>>101496198
>>
File: sdxl_16.jpg (229 KB, 1480x1328)
229 KB
229 KB JPG
Sdg seems pretty slow today
>>
>>101468672
>subject: /sdg/
>body: /ldg/
hm
>>
>>101496462
Baker put in too much salt
>>
File: NOSD3.png (932 KB, 1344x768)
932 KB
932 KB PNG
>>
File: 1721505245984.jpg (90 KB, 768x1152)
90 KB
90 KB JPG
>>
>>101496198
>>101496336
model name?
>>
File: 1720357867447.jpg (84 KB, 768x1152)
84 KB
84 KB JPG
>>
>>101496442
Slow Saturday
>>
>0HJGSG
(You), captchabot.
>>
whats the hardest thing to gen
>>
whats the best way to go about generating two people in comfyui? i remember that a111 i believe had that block/couple thing, is there one for comfy?
>>
>>
>>101497682
need this with Resident Evil girls
>>
>>101497365
Whale

>>101497595
It should have the same extension
>>
>>
Cascade > Pixart ?
>>
>>101497865
awesome
>>
>>101497908
not sure what it is tho, something slimy and wet
>>
uhm
>>
>>101497924
it's a boss monster who's going to try and squish her like that girl in Akira
>>
>>101497883
nah
>>
mmmppfh
>>
>>101496598
pretty sure its pony realism
>>
File: tmpzan22dvw.png (1.34 MB, 896x1152)
1.34 MB
1.34 MB PNG
>>
are there any good generic male pony models? the addetailer keeps turning the generic face feminine
>>
File: 1721505531775.jpg (77 KB, 768x1152)
77 KB
77 KB JPG
>>
File: tmp2j5l794n.png (1.15 MB, 896x1152)
1.15 MB
1.15 MB PNG
>>
File: 1721329584931.png (1.21 MB, 1024x1024)
1.21 MB
1.21 MB PNG
>>
File: tmpykmc7n_i.png (1.09 MB, 896x1152)
1.09 MB
1.09 MB PNG
Somehow I'm struggling to get "bottomless" here
>>
>>101383581
>Asking him to remove the censored pictures so that the model won't see a fucking fat cat everytime a controvertial prompt appears is trolling? The fuck is wrong with you retard?
>>101383593
>That is horseshit, it certainly does not do that, are you retarded?

>>101384331
Auraflow is worthless
>>
>>101498573
I have had limited luck with the OtherStyle_04_maleStyleV2 lora. Most of the pony male model/loras seem to have large breasts. You might be out of luck.

>>101498719
needing to necro that hard. Kinda sad anon.
>>
>>101498900
eight days is far from the furthest anon has gone
>>
File: 000000_15150_.png (2.03 MB, 1060x1413)
2.03 MB
2.03 MB PNG
>>
I might missing something in the A1111/forge APIs. When using "scripts" it isn't a dictionary it is an ordered list, right? It is painful to line everything up and I hate it and feel like I am just missing something.
>>
>>101499098
Very nice
>>
File: 000000_15151_.png (2.01 MB, 1075x1434)
2.01 MB
2.01 MB PNG
>>101499189
TY, red riding hood.
Genned with Kolors, then SD3 USDUpscaled.

>trying for older..
>>
File: 000000_15154_.png (1.89 MB, 1060x1413)
1.89 MB
1.89 MB PNG
>okay moving on, had to remove "little" from prompt.
>>
>>
File: 00000_15155_.png (1.91 MB, 1060x1413)
1.91 MB
1.91 MB PNG
>>
>>101497883
yah
>>
Are there any recommended tools that could turn an image I made with local diffusion into a GIF?
>>
>>101500362
imagemagick
>>
File: ikhjk.jpg (841 KB, 2688x1792)
841 KB
841 KB JPG
OwO

Wut dis?

https://huggingface.co/ostris/16ch-VAE-Adapters
>>
>>101500371
Would it be a manual process? I don't really have any experience with stuff like that. I was hoping for something like an img2gif AI tool that I could just give commands to
>>
>>101500411
fink thats an older version. related stuff here https://huggingface.co/ostris/vae-kl-f8-d16
newer version with comfy nodes soon (they tell me)
>>
>>101500418
>img2gif
Locally? No.
Check OP for guides.
>>
>>101468672
Riffraff free thread
>>
File: tmpyoonkvhi.png (856 KB, 768x768)
856 KB
856 KB PNG
>>
File: __00509_.png (3.19 MB, 2280x1560)
3.19 MB
3.19 MB PNG
>>101490632
4:3 pizza wallpaper
>>
>>101500945
>4:3
3:2
>>
File: tmpva2tkg7j.png (504 KB, 768x768)
504 KB
504 KB PNG
>>
>>101489005
Show me your synthetic model
>>
File: tmp5q_4lz7x.png (665 KB, 768x768)
665 KB
665 KB PNG
>>
File: 00099-471994988.png (3.09 MB, 1344x1728)
3.09 MB
3.09 MB PNG
>>
File: tmpmqhd15a4.png (852 KB, 768x1024)
852 KB
852 KB PNG
>>
File: 06566-269139643-1_8_4.png (2.21 MB, 1224x1224)
2.21 MB
2.21 MB PNG
>>
>>101501355
hes a big guy
>>
File: tmpqi7n92i6.png (1021 KB, 768x768)
1021 KB
1021 KB PNG
>>
File: sdxl_01.jpg (272 KB, 1480x1328)
272 KB
272 KB JPG
>>
>>
File: tmp0z3p15gi.png (1005 KB, 768x1024)
1005 KB
1005 KB PNG
>>
File: tmpekmw40m2.png (636 KB, 768x1024)
636 KB
636 KB PNG
>>
File: ComfyUI_00026_.jpg (645 KB, 1200x1754)
645 KB
645 KB JPG
https://civitai.com/models/570064/b-lora-jean-michel-basquiat-style
>b-lora
huh
>>
File: ComfyUI_00028_.jpg (686 KB, 1200x1754)
686 KB
686 KB JPG
>>101502115
>>
>>
>>
(Character Sheet), (Multiple Poses), (Character Art) are fun ones.
>>
File: 00001-3602996988.png (2.93 MB, 1728x1344)
2.93 MB
2.93 MB PNG
>>101502115
i thought it would have a greater impact on the image composition. I'll play with it more. I feel like this kind of lora does wonders sometimes, I have a jackson pollok lora that is pure kino
>>
File: 00095-3260981830.jpg (286 KB, 2200x1505)
286 KB
286 KB JPG
>>101502373
some of these look really great https://civitai.com/models/246932/bioshock
>jackson pollok lora
link or any examples?
>>
File: __00563_.png (3.28 MB, 1680x2160)
3.28 MB
3.28 MB PNG
>>
File: 00118-538154142.png (2.92 MB, 1280x1920)
2.92 MB
2.92 MB PNG
>>101502432
it's 1.5,https://civitai.com/models/181433/artstyle-jackson-pollock
also this one: https://civitai.com/models/335622/when-ink-meets-watercolor

Image is one I really liked. I find it stunning how abstract artists when trained in lora can output pure fucking kino so often. Really opened my mind on image composition and what it really means.
>>
>>101502476
oh i thought you meant a jackson pollock b-lora. b-loras look to be quicker to train than regular loras but i havent heard much about them
thats a nice image desu
>>
>>101502476
>>101502497
https://github.com/liusida/ComfyUI-B-LoRA
>>
>>101502497
really drunk and didn't notice the b-lora text. Dunno what that even is.
>>
File: 1713739703092141.jpg (171 KB, 712x1152)
171 KB
171 KB JPG
>>
File: __00568_.png (3.03 MB, 1680x2160)
3.03 MB
3.03 MB PNG
>>
File: 116775249263549013-SD.png (3.36 MB, 1408x1656)
3.36 MB
3.36 MB PNG
>>101496598
yep
>>
bigma...
>>
/ldg/ is dead?
>>
>>101501207
Midjourney V6
>>
File: 000000_15168_.png (2.24 MB, 1082x1581)
2.24 MB
2.24 MB PNG
G'mornin Anons, sloooow thread
>>
File: 000000_15173_.png (2.84 MB, 1581x1082)
2.84 MB
2.84 MB PNG
>>
File: Sigma_06172_.png (600 KB, 1280x768)
600 KB
600 KB PNG
>>101504921
Good morning. Can't gen while training. The local anon curse
>>
>>101505623
Nice.

ahh, I just about delved into b-lora training. maybe try that later, need to go get some groceries.......bbiab
>>
File: Sigma_06178_.png (2.08 MB, 1280x768)
2.08 MB
2.08 MB PNG
>>
File: Sigma_06193_.png (1.74 MB, 1280x768)
1.74 MB
1.74 MB PNG
>>
File: Sigma_06214_.png (1.63 MB, 1280x768)
1.63 MB
1.63 MB PNG
>>
File: Sigma_06219_.png (1.79 MB, 1280x768)
1.79 MB
1.79 MB PNG
>>
File: sdxl_13.jpg (255 KB, 1480x1328)
255 KB
255 KB JPG
This thread is 2 days old
>>
File: Sigma_06238_.png (1.24 MB, 1280x768)
1.24 MB
1.24 MB PNG
>>
File: Sigma_06251_.png (1.73 MB, 1280x768)
1.73 MB
1.73 MB PNG
>>101505894
Yup
>>
File: Sigma_06261_.png (1.28 MB, 1280x768)
1.28 MB
1.28 MB PNG
>>
File: Sigma_06270_.png (1.27 MB, 1280x768)
1.27 MB
1.27 MB PNG
>>
File: 000000_15177_.png (2.67 MB, 1459x998)
2.67 MB
2.67 MB PNG
>Buying a few things in Canada always ends up to be $60..F*TRUDUH
>>101505981
>>
File: Sigma_06323_.png (1.03 MB, 1280x768)
1.03 MB
1.03 MB PNG
>>
File: Sigma_06334_.png (343 KB, 1280x768)
343 KB
343 KB PNG
>>101506190
Even the US has experienced massive inflation. At least we don't live in Gaza
>>
>>101506258
Good point.. (not fighting /genociding for Israel either)
>>
>>101506258
>>101506386
Your IP address count has been updated in multiple databases.
>>
>>101506440
I do not consent to my IP address being numbered.
>>
File: 000000_15178_.png (1.97 MB, 1435x982)
1.97 MB
1.97 MB PNG
>>
File: tmp5_3y0144.png (922 KB, 768x1024)
922 KB
922 KB PNG
>>
>>
File: tmpx5yeicyv.png (1012 KB, 768x1024)
1012 KB
1012 KB PNG
>>
File: 0.jpg (152 KB, 1024x1024)
152 KB
152 KB JPG
>>
File: 000000_15180_.png (2.84 MB, 1508x1032)
2.84 MB
2.84 MB PNG
>>
File: 0.jpg (382 KB, 1024x1024)
382 KB
382 KB JPG
>>
File: Sigma_06431_.png (833 KB, 1280x768)
833 KB
833 KB PNG
>>101506772
Clean
>>
File: 0002.jpg (322 KB, 1400x1400)
322 KB
322 KB JPG
I think frequently about the Pringles commercial from.the 90s that was made in semi-claymation style and was to the tune of "gunna make you sweat".
>>
File: Sigma_06469_.png (1.41 MB, 832x1216)
1.41 MB
1.41 MB PNG
>>
>>101506816
>I love this Kolors model.. It's good.
Nice
>>
>>
File: tmp_y1i069x.png (1.02 MB, 768x768)
1.02 MB
1.02 MB PNG
>>
File: Sigma_06532_.png (1.85 MB, 1280x768)
1.85 MB
1.85 MB PNG
>>
why isn't forge in the recommended UI list, is there something wrong with it?
also is it normal that it takes me like 15 min to load checkpoints?(gtx 1070)
>>
File: tmpsirfvja_.png (921 KB, 768x1024)
921 KB
921 KB PNG
It's funny how without score tags, Pony's output sometimes looks like visual novel CG
>>
File: Sigma_06539_.png (1.31 MB, 832x1216)
1.31 MB
1.31 MB PNG
>>101506996
afaik they went experimental and broke most existing extensions
>>
>>101506996
forge is in the process of major changes that will break compatibility with most of its plugins
it still works fine as of now but the devs have said if you want to keep it running as is, then never update beyond the current version
it's easier to just steer people away at this point
>>
>>
File: 000000_15182_.png (2.44 MB, 1581x1082)
2.44 MB
2.44 MB PNG
>>
>>101506996
>also is it normal that it takes me like 15 min to load checkpoints?(gtx 1070)
that doesn't sound normal, how much ram do you have? forge will load a checkpoint into ram and then keep it there until it needs to use a different one. if your ram is full, it's likely throwing the model into the swap space and thrashing the disk instead of loading it directly from ram to vram.
oh and if you have ample ram look in settings, you can actually have forge keep multiple checkpoints in memory at once so they can be swapped over to the gpu more quickly
>>
>>101507167
16gb.
does it running from a slow HDD can have something to do with it?
will moving everything onto an SSD can help?
>>
>>101507316
>will moving everything onto an SSD can help?
Yes, ideally an Nvme..
Adding another 16Gb of Ram will. I was having problems even with 32, not with 64.
>>
File: Sigma_06570_.png (1.29 MB, 1280x768)
1.29 MB
1.29 MB PNG
>>
File: 000000_15185_.png (2.04 MB, 1581x1082)
2.04 MB
2.04 MB PNG
>>
File: comfyui.jpg (179 KB, 2098x1020)
179 KB
179 KB JPG
Why have you still not upgraded to ComfyUI, the white man's solution to local diffusion?
Sure, the learning curve is almost a cliff, but once you are in the flow every lesser solution seems like undercooked madness.

Besides the actual image generation, it's also 10 times faster than WebUI.
>>
File: tmptsnn9_pd.png (914 KB, 768x768)
914 KB
914 KB PNG
>>
File: Sigma_06602_.png (1.1 MB, 1280x768)
1.1 MB
1.1 MB PNG
>>101507464
1click for 1girl
>>
File: 000000_15186_.png (2.56 MB, 1581x1082)
2.56 MB
2.56 MB PNG
>>
Fresh bread, straight outta oven...


>>101507621
>>101507621
>>101507621
>>
>>101507647
You're a few images too early

Is this how that news guy always gets the jump?
>>
File: 000000_15188_.png (3.24 MB, 1082x1581)
3.24 MB
3.24 MB PNG
>>
>>101508250
nice
>>
>>101504914
>Source needed
>>
File: 000000_15192_.png (1.92 MB, 1003x1467)
1.92 MB
1.92 MB PNG
>>101508781
>looks like nick cage to me..
fill the thread up!!ppl kek
>>
File: __00580_.png (1.4 MB, 896x1152)
1.4 MB
1.4 MB PNG
>>
File: __00581_.png (1.2 MB, 896x1152)
1.2 MB
1.2 MB PNG
>>
File: __00584_.png (1.24 MB, 896x1152)
1.24 MB
1.24 MB PNG
>>
File: 1718546602852953.jpg (227 KB, 1200x1754)
227 KB
227 KB JPG



[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.