/ldg/ - Local Diffusion General
Anonymous 01/12/25(Sun)22:15:45 | 345 comments | 77 images | 🔒 Locked
Discussion of Free and Open-Source Diffusion models.
Human Consciousness Edition
Last Time: >>103870533
Newfrien Guide: https://rentry.org/ldg-lazy-getting-started-guide
>UI
Metastable:https://metastable.studio
SwarmUI:https://github.com/mcmonkeyprojects/SwarmUI
Forge:https://github.com/lllyasviel/stable-diffusion-webui-forge
reForge:https://github.com/Panchovix/stable-diffusion-webui-reForge
ComfyUI:https://github.com/comfyanonymous/ComfyUI
InvokeAI:https://github.com/invoke-ai/InvokeAI
>Models, LoRAs, & Upscalers
https://civitai.com
https://tensor.art/
https://openmodeldb.info
>Training
https://github.com/Nerogar/OneTrainer
https://github.com/kohya-ss/sd-scripts/tree/sd3
https://github.com/derrian-distro/LoRA_Easy_Training_Scripts
>HunyuanVideo
Comfy:https://github.com/kijai/ComfyUI-HunyuanVideoWrapper/
Windows:https://rentry.org/crhcqq54
Training:https://github.com/tdrussell/diffusion-pipe|https://github.com/kohya-ss/musubi-tuner
>Flux
Forge:https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/1050
ComfyUI:https://comfyanonymous.github.io/ComfyUI_examples/flux
DeDistilled:https://huggingface.co/TheYuriLover/flux-dev-de-distill-GGUF/tree/main
Pixelwave:https://civitai.com/models/141592/pixelwave
>Misc
Share Metadata:https://catbox.moe|https://litterbox.catbox.moe/
Img2Prompt:https://huggingface.co/spaces/fancyfeast/joy-caption-alpha-two
Archive:https://rentry.org/sdg-link
Samplers:https://stable-diffusion-art.com/samplers/
Digital Art Software:https://krita.org/en/
Txt2Img Plugin:https://kritaaidiffusion.com/
Collagebaker:https://www.befunky.com/create/collage/
Video Collagebaker:https://kdenlive.org/en/
>Related
>>>/aco/csdg
>>>/b/degen
>>>/c/kdg
>>>/d/ddg
>>>/e/edg
>>>/h/hdg
>>>/trash/slop
>>>/vt/vtai
>Local Sexting
>>>/g/lmg
>Maintain thread quality
https://rentry.org/debo
Human Consciousness Edition
Last Time: >>103870533
Newfrien Guide: https://rentry.org/ldg-lazy-getting
>UI
Metastable:https://metastable.studi
SwarmUI:https://github.com/mcmonkey
Forge:https://github.com/lllyasviel
reForge:https://github.com/Panchovi
ComfyUI:https://github.com/comfyano
InvokeAI:https://github.com/invoke-
>Models, LoRAs, & Upscalers
https://civitai.com
https://tensor.art/
https://openmodeldb.info
>Training
https://github.com/Nerogar/OneTrain
https://github.com/kohya-ss/sd-scri
https://github.com/derrian-distro/L
>HunyuanVideo
Comfy:https://github.com/kijai/Comf
Windows:https://rentry.org/crhcqq54
Training:https://github.com/tdrusse
>Flux
Forge:https://github.com/lllyasviel
ComfyUI:https://comfyanonymous.gith
DeDistilled:https://huggingface.co/
Pixelwave:https://civitai.com/model
>Misc
Share Metadata:https://catbox.moe|https:/
Img2Prompt:https://huggingface.co/s
Archive:https://rentry.org/sdg-link
Samplers:https://stable-diffusion-a
Digital Art Software:https://krita.org/en/
Txt2Img Plugin:https://kritaaidiffusion.com
Collagebaker:https://www.befunky.co
Video Collagebaker:https://kdenlive.org/e
>Related
>>>/aco/csdg
>>>/b/degen
>>>/c/kdg
>>>/d/ddg
>>>/e/edg
>>>/h/hdg
>>>/trash/slop
>>>/vt/vtai
>Local Sexting
>>>/g/lmg
>Maintain thread quality
https://rentry.org/debo
Anonymous 01/12/25(Sun)22:17:23 No.103873341
>>103873329
Thanks, vishnu monkey.
Thanks, vishnu monkey.
Anonymous 01/12/25(Sun)22:18:03 No.103873348
>>103873329
damn, we get new threads quickly now, I kind miss the time a single /ldg/ thread could last 2 days lol
damn, we get new threads quickly now, I kind miss the time a single /ldg/ thread could last 2 days lol
Anonymous 01/12/25(Sun)22:23:40 No.103873380
>average /ldg/ poster
Anonymous 01/12/25(Sun)22:24:41 No.103873389
>>103873380
I'm above average so I'm good
I'm above average so I'm good
Anonymous 01/12/25(Sun)22:24:58 No.103873393
Anonymous 01/12/25(Sun)22:34:24 No.103873467
>>103873329
>two bukkake spam asian girl plus return to monke, catjak proportion girl, sigma balls graph (no one gives a shit), 0 anon fishing for a reply for his Sisyphus Happy reddit meme (which is kind of iconic because this is actually kind of actually art, because it depicts him gooning out loud as he pushes his shitty gens up a hill forever, getting no bitches ever, as he cries out in despair and goon anger)
asking again, why is it when I search sdg in the catalog it refers me to this shithole?
>two bukkake spam asian girl plus return to monke, catjak proportion girl, sigma balls graph (no one gives a shit), 0 anon fishing for a reply for his Sisyphus Happy reddit meme (which is kind of iconic because this is actually kind of actually art, because it depicts him gooning out loud as he pushes his shitty gens up a hill forever, getting no bitches ever, as he cries out in despair and goon anger)
asking again, why is it when I search sdg in the catalog it refers me to this shithole?
Anonymous 01/12/25(Sun)22:34:51 No.103873473
Anonymous 01/12/25(Sun)22:36:03 No.103873478
>>103873389
The better looking you are IRL the better your gens are. This is a fact.
The better looking you are IRL the better your gens are. This is a fact.
Anonymous 01/12/25(Sun)22:36:47 No.103873482
maskeditor is broken now, do these people test anything before pushing their commits? the incompetence of this "comfy org" team is baffling
Anonymous 01/12/25(Sun)22:37:31 No.103873486
Anonymous 01/12/25(Sun)22:38:12 No.103873497
Anonymous 01/12/25(Sun)22:40:05 No.103873507
>>103873329
for the love of god, more mayli
for the love of god, more mayli
Anonymous 01/12/25(Sun)22:40:33 No.103873510
>>103873393
here you go
here you go
Anonymous 01/12/25(Sun)22:40:39 No.103873512
mayli thread
Anonymous 01/12/25(Sun)22:44:10 No.103873543
>>103873467
At least the Mayli poster doesn't actually pretend to be Mayli.
At least the Mayli poster doesn't actually pretend to be Mayli.
Anonymous 01/12/25(Sun)22:47:19 No.103873570
You ever type out a cheeky response to obvious b8 only to delete it and not post?
Anonymous 01/12/25(Sun)22:47:34 No.103873572
>>103873510
all that changed was the seed.
all that changed was the seed.
Anonymous 01/12/25(Sun)22:48:41 No.103873581
Serious question.
Is there a game where your character is an old ugly woman?
Is there a game where your character is an old ugly woman?
Anonymous 01/12/25(Sun)22:48:55 No.103873589
/sdg/ is like eating a decent steak with your choice of temperature with decent service but the waiter drones on and takes too long but still delivers a hot meal eventually
/ldg/ is like getting served a well done steak even if you didn't ask for it and the waiter yells at you for not picking the correct wine to go with the steak but he won't answer what wine you were supposed to pick nor did he ask how well you wanted your steak in the first place,
/ldg/ is like getting served a well done steak even if you didn't ask for it and the waiter yells at you for not picking the correct wine to go with the steak but he won't answer what wine you were supposed to pick nor did he ask how well you wanted your steak in the first place,
Anonymous 01/12/25(Sun)22:49:23 No.103873595
>>103873570
yeah that happens when I'm not in the mood of arguing with retards for the day lol
yeah that happens when I'm not in the mood of arguing with retards for the day lol
Anonymous 01/12/25(Sun)22:49:32 No.103873597
>>103873570
I would never admit this
I would never admit this
Anonymous 01/12/25(Sun)22:49:59 No.103873601
>>103873486
whats funny to me is comfy saying that there are people making six figures by using comfyui for their work, lmao and this is the shitty repo that they present
whats funny to me is comfy saying that there are people making six figures by using comfyui for their work, lmao and this is the shitty repo that they present
Anonymous 01/12/25(Sun)22:50:00 No.103873602
>>103873597
kekd
kekd
Anonymous 01/12/25(Sun)22:50:25 No.103873608
>>103873589
>and yet he's still lurking at the restaurant he hates instead of eating good on the /sdg/ restaurant
curious
>and yet he's still lurking at the restaurant he hates instead of eating good on the /sdg/ restaurant
curious
Anonymous 01/12/25(Sun)22:57:22 No.103873671
>103873589
Take your insecurity elsewhere
Take your insecurity elsewhere
Anonymous 01/12/25(Sun)22:58:15 No.103873684
Anonymous 01/12/25(Sun)22:58:52 No.103873695
>>103873684
dude I want a furk X mayli crossover on one of your videos kek
dude I want a furk X mayli crossover on one of your videos kek
Anonymous 01/12/25(Sun)22:59:34 No.103873701
>>103873684
That hand is impossible!
That hand is impossible!
Anonymous 01/12/25(Sun)23:03:12 No.103873735
>dont train on my drawings chud
Anonymous 01/12/25(Sun)23:04:56 No.103873748
>>103873735
oh fug im matrix decompoooosing
oh fug im matrix decompoooosing
Anonymous 01/12/25(Sun)23:05:21 No.103873752
>>103873735
that's why I don't care if the artists get a downfall, they are too unobnoxious with their copyright shit, fuck them
that's why I don't care if the artists get a downfall, they are too unobnoxious with their copyright shit, fuck them
Anonymous 01/12/25(Sun)23:05:55 No.103873757
https://www.reddit.com/r/comfyui/comments/1hzsobm/los_angeles_fire_ai_generation_comfyui_3080_rtx/
kek, redditors are something else
kek, redditors are something else
Anonymous 01/12/25(Sun)23:07:39 No.103873768
>>103873757
kek, since when ledditors are more edgy than us?? /pol/ bros, how do we respond?
kek, since when ledditors are more edgy than us?? /pol/ bros, how do we respond?
Anonymous 01/12/25(Sun)23:09:25 No.103873775
Anonymous 01/12/25(Sun)23:12:33 No.103873795
>>103873775
You will never be American.
You will never be American.
Anonymous 01/12/25(Sun)23:15:05 No.103873812
>>103873393
here, I fixed it
here, I fixed it
Anonymous 01/12/25(Sun)23:15:21 No.103873814
>>103873795
Elon disagree with that, prepare for a mass exodus of sirs in murica, murica will be like India 2 (previously named Canada) in less than 10 years kek
Elon disagree with that, prepare for a mass exodus of sirs in murica, murica will be like India 2 (previously named Canada) in less than 10 years kek
Anonymous 01/12/25(Sun)23:17:07 No.103873827
>>103873814
We will have a kiwi-e-mart every few houses, and you say this like it's a bad thing?
We will have a kiwi-e-mart every few houses, and you say this like it's a bad thing?
Anonymous 01/12/25(Sun)23:17:36 No.103873833
>>103873827
kek
kek
Anonymous 01/12/25(Sun)23:18:02 No.103873836
>>103873812
or maybe this one. NAILED IT.
or maybe this one. NAILED IT.
Anonymous 01/12/25(Sun)23:19:01 No.103873848
I am trying the LoRA caption load custom node to try to start training my first LoRA within Comfy, after queuing the processing it printed 'got prompt' to cmd and now it's just sitting at 0%, it's at 50% CPU usage and 35-50gb of ram usage, but nothing is actually printing at all. Should I just wait? I kind of assume that if it crashes it would actually tell me as it's generally always done so so far.
Anonymous 01/12/25(Sun)23:19:03 No.103873849
>>103873836
wait let me fix the hand!
wait let me fix the hand!
Anonymous 01/12/25(Sun)23:21:39 No.103873876
>>103873589
I don't relate to this because I can't afford steak after spending all my money on GPUs
I don't relate to this because I can't afford steak after spending all my money on GPUs
Anonymous 01/12/25(Sun)23:25:08 No.103873903
>>103873848
I don't know about any of that but you're going to have a much easier time training your LoRA in a dedicated repository in its own venv
I don't know about any of that but you're going to have a much easier time training your LoRA in a dedicated repository in its own venv
Anonymous 01/12/25(Sun)23:25:13 No.103873904
>>103873836
there!
there!
Anonymous 01/12/25(Sun)23:26:07 No.103873913
>>103873904
You're disgusting
You're disgusting
Anonymous 01/12/25(Sun)23:26:55 No.103873921
Anonymous 01/12/25(Sun)23:27:42 No.103873929
>>103873904
Hideous
Hideous
Anonymous 01/12/25(Sun)23:27:58 No.103873931
>>103873848
Did you check the console? Also what type of caption is it doing? JoyTag/JoyCaption? Florence? WD14 v3? Something with a big LLM?
Did you check the console? Also what type of caption is it doing? JoyTag/JoyCaption? Florence? WD14 v3? Something with a big LLM?
Anonymous 01/12/25(Sun)23:28:05 No.103873932
>>103873921
That's what I do to my concubine when they bad mouth me.
That's what I do to my concubine when they bad mouth me.
Anonymous 01/12/25(Sun)23:30:14 No.103873952
Anonymous 01/12/25(Sun)23:30:55 No.103873957
i want to kill the Mayli poster
Anonymous 01/12/25(Sun)23:31:24 No.103873960
>>103873929
You've never seen a woman without makeup before?
You've never seen a woman without makeup before?
Anonymous 01/12/25(Sun)23:33:00 No.103873973
>>103873957
https://youtu.be/Tg3d2Hj-RV4?t=79
https://youtu.be/Tg3d2Hj-RV4?t=79
Anonymous 01/12/25(Sun)23:35:36 No.103873989
Anonymous 01/12/25(Sun)23:36:24 No.103873994
>>103873903
I am new to this, hopefully once I have more experience I will know how to do that successfully.
>>103873931
It's WD14 V3 but it was failing in the load step, the console wasn't printing anything. I've started running images in batches and that seemed to fix it, so maybe I was just being impatient.
I am new to this, hopefully once I have more experience I will know how to do that successfully.
>>103873931
It's WD14 V3 but it was failing in the load step, the console wasn't printing anything. I've started running images in batches and that seemed to fix it, so maybe I was just being impatient.
Anonymous 01/12/25(Sun)23:36:38 No.103873996
>>103873929
ok I fixed it
ok I fixed it
Anonymous 01/12/25(Sun)23:37:30 No.103874005
>>103873994
>I am new to this, hopefully once I have more experience I will know how to do that successfully.
No I am telling you. Go to any LoRA trainer repository and on github and follow the steps there because training within comfyui is an overly complicated mess.
>I am new to this, hopefully once I have more experience I will know how to do that successfully.
No I am telling you. Go to any LoRA trainer repository and on github and follow the steps there because training within comfyui is an overly complicated mess.
Anonymous 01/12/25(Sun)23:38:28 No.103874015
>>103874005
I will look into this, you are very right, thank you anon
I will look into this, you are very right, thank you anon
Anonymous 01/12/25(Sun)23:43:26 No.103874057
>>103873952
I love Juice
>>103873989
schizo space angel
>>103873996
I think you're missing something on the shoulder
I love Juice
>>103873989
schizo space angel
>>103873996
I think you're missing something on the shoulder
Anonymous 01/12/25(Sun)23:43:50 No.103874062
Was sorting my gens with a script to divide them by month the file was modified and I noticed something odd, there were NO GENS for the entire month of November in 2023. Did some looking around my pictures folder and realized that in November 2023 I did nothing except Dall-e 3. I gave up local diffusion for an entire month.
Reflecting on this, I feel ashamed.
Reflecting on this, I feel ashamed.
Anonymous 01/12/25(Sun)23:44:33 No.103874066
>>103874062
its okay anon i still love you
its okay anon i still love you
Anonymous 01/12/25(Sun)23:44:46 No.103874067
>>103873996
I got it. She needed blue eyes. Now she is beautiful.
I got it. She needed blue eyes. Now she is beautiful.
Anonymous 01/12/25(Sun)23:45:28 No.103874072
>>103874066
I mean, dalle3 was a revolutionary model, can't blame you on that one anon, 2023 was the year OpenAI destroyed the competition (GPT4 and dalle3) and it wasn't even close
I mean, dalle3 was a revolutionary model, can't blame you on that one anon, 2023 was the year OpenAI destroyed the competition (GPT4 and dalle3) and it wasn't even close
Anonymous 01/12/25(Sun)23:45:41 No.103874073
>>103873957
Many men wish death upon me
shitty gens in LDG
I'm tryin' to be what I'm destined to be
And niggas tryin' to take Mayli away (Come on)
I put a hole in a nigga for fuckin' with me
Mayli on the OP wall, now you gon' see
Better watch how you talk when you talk about me
'Cause I'll come and take your GPU away
Many men wish death upon me
shitty gens in LDG
I'm tryin' to be what I'm destined to be
And niggas tryin' to take Mayli away (Come on)
I put a hole in a nigga for fuckin' with me
Mayli on the OP wall, now you gon' see
Better watch how you talk when you talk about me
'Cause I'll come and take your GPU away
Anonymous 01/12/25(Sun)23:46:27 No.103874081
>>103874073
Thread theme
Thread theme
Anonymous 01/12/25(Sun)23:46:42 No.103874084
Anonymous 01/12/25(Sun)23:47:22 No.103874088
>>103873957
Don't worry, when she finds out that anon is going to get sued the shit out of
Don't worry, when she finds out that anon is going to get sued the shit out of
Anonymous 01/12/25(Sun)23:48:27 No.103874095
>>103874088
would be based to know that Mayli is a /ldg/ lurker kek
would be based to know that Mayli is a /ldg/ lurker kek
Anonymous 01/12/25(Sun)23:49:53 No.103874103
Anonymous 01/12/25(Sun)23:49:58 No.103874104
>>103874088
sued for what exactly?
sued for what exactly?
Anonymous 01/12/25(Sun)23:50:02 No.103874106
>>103874095
People are posting these videos on X in increasing frequency
People are posting these videos on X in increasing frequency
Anonymous 01/12/25(Sun)23:50:26 No.103874110
Anonymous 01/12/25(Sun)23:50:38 No.103874112
>>103874106
really? can you show some links that's interesting
really? can you show some links that's interesting
Anonymous 01/12/25(Sun)23:51:29 No.103874118
>>103874104
She owns the copyright to those videos, so models trained on them and resulting outputs are open to copyright violation. There are also impersonation/deepfake laws that apply even to non-pornographic works.
She owns the copyright to those videos, so models trained on them and resulting outputs are open to copyright violation. There are also impersonation/deepfake laws that apply even to non-pornographic works.
Anonymous 01/12/25(Sun)23:51:57 No.103874125
Get a load of this doomer
Anonymous 01/12/25(Sun)23:52:28 No.103874130
>>103874118
>There are also impersonation/deepfake laws that apply even to non-pornographic works.
so all the guys who make celebrities loras on civitai can be sued?
>There are also impersonation/deepfake laws that apply even to non-pornographic works.
so all the guys who make celebrities loras on civitai can be sued?
Anonymous 01/12/25(Sun)23:53:00 No.103874134
>>103874057
It's complete now. Thank you for your guidance.
It's complete now. Thank you for your guidance.
Anonymous 01/12/25(Sun)23:54:17 No.103874145
>>103873994
> maybe I was just being impatient.
Could be.
>>103874005
The flux trainer in comfyui was really quite easy. But sure, there are other options.
> maybe I was just being impatient.
Could be.
>>103874005
The flux trainer in comfyui was really quite easy. But sure, there are other options.
Anonymous 01/12/25(Sun)23:54:18 No.103874146
>But the pedo poster keeps getting away with it.
Anonymous 01/12/25(Sun)23:54:27 No.103874147
>>103874130
If Trump puts another stupid bimbo on the courts, it will be illegal to own a penis.
If Trump puts another stupid bimbo on the courts, it will be illegal to own a penis.
Anonymous 01/12/25(Sun)23:55:02 No.103874151
>>103874118
>She owns the copyright to those videos
that's assuming he trained the lora on those videos, recently she does art and shit so there's videos of her that are just regular videos, there's not just the 3 porn videos of her actually
>She owns the copyright to those videos
that's assuming he trained the lora on those videos, recently she does art and shit so there's videos of her that are just regular videos, there's not just the 3 porn videos of her actually
Anonymous 01/12/25(Sun)23:55:05 No.103874153
>>103874147
thankfully i already cut mine off before i installed comfyui
thankfully i already cut mine off before i installed comfyui
Anonymous 01/12/25(Sun)23:56:03 No.103874162
>>103874153
kek, dunno why the troon comfyui meme appeared though, comfy is a regular dude he's not some mentaly ill troon
kek, dunno why the troon comfyui meme appeared though, comfy is a regular dude he's not some mentaly ill troon
Anonymous 01/12/25(Sun)23:56:44 No.103874169
>>103874153
Can't trick me, I know you haven't sold it.
Can't trick me, I know you haven't sold it.
Anonymous 01/12/25(Sun)23:57:09 No.103874173
>>103874118
image outputs cannot be copyrighted anon
image outputs cannot be copyrighted anon
Anonymous 01/12/25(Sun)23:58:13 No.103874179
>>103874173
*yet
*yet
Anonymous 01/12/25(Sun)23:58:26 No.103874180
>>103874118
> there are also impersonation/deepfake laws that apply even to non-pornographic works.
depends on the nation, people who put themselves into the public as actor or politician or w/e do often not have protection
> so models trained on them and resulting outputs are open to copyright violation
AFAIK not a single country has ruled that learning from something is a copyright violation and most likely won't since it's the same as huamns artists learning from copyrighted works which is also not covered by copyright
it's just what the anti AI crowd WISHED would happen
> there are also impersonation/deepfake laws that apply even to non-pornographic works.
depends on the nation, people who put themselves into the public as actor or politician or w/e do often not have protection
> so models trained on them and resulting outputs are open to copyright violation
AFAIK not a single country has ruled that learning from something is a copyright violation and most likely won't since it's the same as huamns artists learning from copyrighted works which is also not covered by copyright
it's just what the anti AI crowd WISHED would happen
Anonymous 01/12/25(Sun)23:59:04 No.103874184
>>103874173
there's a reason why every company shit themselves when we ask them if they trained their model on copyrighted shit, they know they're in big trouble if we found out
https://youtu.be/mAUpxN-EIgU?t=263
there's a reason why every company shit themselves when we ask them if they trained their model on copyrighted shit, they know they're in big trouble if we found out
https://youtu.be/mAUpxN-EIgU?t=263
Anonymous 01/12/25(Sun)23:59:18 No.103874188
>>103874173
no but likeness can be sued over if it's beyond a reasonable doubt.
no but likeness can be sued over if it's beyond a reasonable doubt.
Anonymous 01/13/25(Mon)00:00:05 No.103874193
>>103874188
you know a video model is great when you could be sued for it, no one is gonna sue some random doing a Mayli lora on pyramid flow kek
you know a video model is great when you could be sued for it, no one is gonna sue some random doing a Mayli lora on pyramid flow kek
Anonymous 01/13/25(Mon)00:00:09 No.103874194
>>103874180
>AFAIK not a single country has ruled that learning from something is a copyright violation
Don't ask for a source but Japan ruled in favor of violation but only because it depicted a specific character. You're still correct however.
>AFAIK not a single country has ruled that learning from something is a copyright violation
Don't ask for a source but Japan ruled in favor of violation but only because it depicted a specific character. You're still correct however.
Anonymous 01/13/25(Mon)00:00:19 No.103874196
im snitching
Anonymous 01/13/25(Mon)00:01:09 No.103874206
>>103874193
Nah nobody is going to sue it's too difficult to track down and find who to even blame for it.
Nah nobody is going to sue it's too difficult to track down and find who to even blame for it.
Anonymous 01/13/25(Mon)00:01:33 No.103874211
>>103874057
"schizophrenic" is in the negatives, always. if you can believe it
"schizophrenic" is in the negatives, always. if you can believe it
Anonymous 01/13/25(Mon)00:02:45 No.103874217
>>103874206
desu I expected much more lawsuit during the AI era, it's been more than 2 years it bloomed and so far the only thing I remember is the New York Times sueing OpenAI, but that's pretty much it?
desu I expected much more lawsuit during the AI era, it's been more than 2 years it bloomed and so far the only thing I remember is the New York Times sueing OpenAI, but that's pretty much it?
Anonymous 01/13/25(Mon)00:02:55 No.103874218
>>103874173
If they're generated from an overtrained model that obviously infringes on a copyright work, then yes, they can.
If they're generated from an overtrained model that obviously infringes on a copyright work, then yes, they can.
Anonymous 01/13/25(Mon)00:03:02 No.103874220
>>103874184
>there's a reason why every company shit themselves when we ask them if they trained their model on copyrighted shit
trained on copyrighted material yes that is trouble, the outputs you generate cannot be copyrighted
>>103874188
oh really? lets sue impersonators and doubles then, lets prohibit comedy sketches that use makeup. Anon, stop making laws in your head, you sound like Emad
>there's a reason why every company shit themselves when we ask them if they trained their model on copyrighted shit
trained on copyrighted material yes that is trouble, the outputs you generate cannot be copyrighted
>>103874188
oh really? lets sue impersonators and doubles then, lets prohibit comedy sketches that use makeup. Anon, stop making laws in your head, you sound like Emad
Anonymous 01/13/25(Mon)00:03:05 No.103874221
>>103874184
it's not a bad thing that they train on copyrighted shit
copyright is a joke and should always be ignored, or even better, actively disrespected and violated, especially if something useful and practical can come from doing so
it's not a bad thing that they train on copyrighted shit
copyright is a joke and should always be ignored, or even better, actively disrespected and violated, especially if something useful and practical can come from doing so
Anonymous 01/13/25(Mon)00:03:45 No.103874230
>>103874218
Where exactly do you draw the distinction here?
Where exactly do you draw the distinction here?
Anonymous 01/13/25(Mon)00:04:30 No.103874234
>>103874220
>trained on copyrighted material yes that is trouble
which is the topic at hand, since Mayli has the copyright of the porn video, technically you can't really train on those videos
>trained on copyrighted material yes that is trouble
which is the topic at hand, since Mayli has the copyright of the porn video, technically you can't really train on those videos
Anonymous 01/13/25(Mon)00:04:35 No.103874236
>>103874194
ai or not, you cannot publish mickey mouse in most countries, but maybe mickey mouse parody or porn spoofs are allowed... or not
ai or not, you cannot publish mickey mouse in most countries, but maybe mickey mouse parody or porn spoofs are allowed... or not
Anonymous 01/13/25(Mon)00:05:01 No.103874239
They try to bring Mayli poster down....
Many men....
Many men....
Anonymous 01/13/25(Mon)00:05:06 No.103874240
Anonymous 01/13/25(Mon)00:05:24 No.103874243
>>103874234
yeah sue the people who trained the lora then, not the one anon generating parody images, thats free speech
yeah sue the people who trained the lora then, not the one anon generating parody images, thats free speech
Anonymous 01/13/25(Mon)00:05:34 No.103874244
>>103874221
>copyright is a joke and should always be ignored, or even better, actively disrespected and violated
I 100% agree with that, copyright is for me one of the worst poison of humanity, it fucking kills the fun and the creativity, great artists steal after all
>copyright is a joke and should always be ignored, or even better, actively disrespected and violated
I 100% agree with that, copyright is for me one of the worst poison of humanity, it fucking kills the fun and the creativity, great artists steal after all
Anonymous 01/13/25(Mon)00:06:18 No.103874248
>>103874234
What if the model is trained on the output of other outputs?
What if the model is trained on the output of other outputs?
Anonymous 01/13/25(Mon)00:06:36 No.103874251
>>103874243
>sue the people who trained the lora then, not the one anon generating parody images
that's the same guy though? he has that lora for himself he never posted it on the internet
>sue the people who trained the lora then, not the one anon generating parody images
that's the same guy though? he has that lora for himself he never posted it on the internet
Anonymous 01/13/25(Mon)00:06:43 No.103874254
>>103874220
>In most US states, you can sue someone for using your name or likeness without your consent in an exploitative way.
This excludes "transformative art" like portraits, parody, and such. Porn is 100% exploitative. Morally I don't give a fuck just know tay tay is gonna sue someone.
>In most US states, you can sue someone for using your name or likeness without your consent in an exploitative way.
This excludes "transformative art" like portraits, parody, and such. Porn is 100% exploitative. Morally I don't give a fuck just know tay tay is gonna sue someone.
Anonymous 01/13/25(Mon)00:06:59 No.103874255
>>103874218
no anon, you cannot copyright ai generated content, therefore you cannot sue for that, generated content =/= copyrighted material is not the same, you cannot reproduce exact copies of something using AI
no anon, you cannot copyright ai generated content, therefore you cannot sue for that, generated content =/= copyrighted material is not the same, you cannot reproduce exact copies of something using AI
Anonymous 01/13/25(Mon)00:07:50 No.103874259
>>103874255
>you cannot reproduce exact copies of something using AI
what if one day AI is so good it can reproduce a fucking movie?
>you cannot reproduce exact copies of something using AI
what if one day AI is so good it can reproduce a fucking movie?
Anonymous 01/13/25(Mon)00:08:05 No.103874262
Gens in my eyes and I can't see....
Anonymous 01/13/25(Mon)00:09:18 No.103874266
>>103874220 >>103874234
>trained on copyrighted material yes that is trouble
not a fact in most countries, may never be a fact - again artists are permitted to learn from copyrighted materials and so is probably ai
there are people who don't like it and who would like to have exclusive ownership over every idea wielding copyright like a much broader patent with no expiration, screw them
>trained on copyrighted material yes that is trouble
not a fact in most countries, may never be a fact - again artists are permitted to learn from copyrighted materials and so is probably ai
there are people who don't like it and who would like to have exclusive ownership over every idea wielding copyright like a much broader patent with no expiration, screw them
Anonymous 01/13/25(Mon)00:10:09 No.103874271
>>103874118
>implying she would be mad
I talked to her on the phone and she said its fine, actually she likes it
>implying she would be mad
I talked to her on the phone and she said its fine, actually she likes it
Anonymous 01/13/25(Mon)00:10:10 No.103874272
Anonymous 01/13/25(Mon)00:10:19 No.103874273
>>103874266
>artists are permitted to learn from copyrighted materials and so is probably ai
I agree with that, that's logical and intuitive, but we're talking about laws here, not common sense
>artists are permitted to learn from copyrighted materials and so is probably ai
I agree with that, that's logical and intuitive, but we're talking about laws here, not common sense
Anonymous 01/13/25(Mon)00:10:44 No.103874276
>>103874254
the Mayli poster of /ldg/ hasn't posted anything porn-related, just Mayli in funny situations, parody portraits and stuff like that
the Mayli poster of /ldg/ hasn't posted anything porn-related, just Mayli in funny situations, parody portraits and stuff like that
Anonymous 01/13/25(Mon)00:11:35 No.103874284
>>103874255
you probably automatically have copyright on your ai generated content
>>103874273
the laws say nothing else so far, duh
maybe eventually some anti ai country will fuck up their laws, but it likely won't catch on as said country's companies will likely then outsource a shitload of work to china or w/e instead
you probably automatically have copyright on your ai generated content
>>103874273
the laws say nothing else so far, duh
maybe eventually some anti ai country will fuck up their laws, but it likely won't catch on as said country's companies will likely then outsource a shitload of work to china or w/e instead
Anonymous 01/13/25(Mon)00:12:15 No.103874288
>>103874259
reproduce how exactly? like copying and pasting a file? then why use AI for that exactly lmao
reproduce how exactly? like copying and pasting a file? then why use AI for that exactly lmao
Anonymous 01/13/25(Mon)00:13:34 No.103874302
Anonymous 01/13/25(Mon)00:13:53 No.103874305
>>103874284
>the laws say nothing else so far, duh
there's a reason for that, China don't give a fuck about copyright, and the US don't want to let them eat all the cake, god I love competition
>>103874288
>then why use AI for that exactly lmao
it's not like it's a goal to get a "copy paste" AI, but ultimately a perfect AI model is able to reproduce the training data with perfection, which is what we want actually, I'm not about overtraining an AI, but like an AI who's just good at learning stuff
>the laws say nothing else so far, duh
there's a reason for that, China don't give a fuck about copyright, and the US don't want to let them eat all the cake, god I love competition
>>103874288
>then why use AI for that exactly lmao
it's not like it's a goal to get a "copy paste" AI, but ultimately a perfect AI model is able to reproduce the training data with perfection, which is what we want actually, I'm not about overtraining an AI, but like an AI who's just good at learning stuff
Anonymous 01/13/25(Mon)00:14:54 No.103874314
https://github.com/facok/ComfyUI-HunyuanVideoMultiLora
is this a meme or does it really fix the issue you have when going for multiple hunyuan loras?
is this a meme or does it really fix the issue you have when going for multiple hunyuan loras?
Anonymous 01/13/25(Mon)00:16:08 No.103874323
>>103874284
Nope, you cant copyright ai generated content, there have been many tries of people trying to copyright their AI generated crap without success
https://www.reuters.com/legal/litigation/artist-sues-after-us-rejects-copyright-ai-generated-image-2024-09-26/
Nope, you cant copyright ai generated content, there have been many tries of people trying to copyright their AI generated crap without success
https://www.reuters.com/legal/litig
Anonymous 01/13/25(Mon)00:16:24 No.103874325
>>103874314
Haven't tried it yet.
Haven't tried it yet.
Anonymous 01/13/25(Mon)00:17:00 No.103874331
>>103874323
so basically if an AI fag puts his AI work on a patreon, we can pirate his shit and put them on the internet without much trouble?
so basically if an AI fag puts his AI work on a patreon, we can pirate his shit and put them on the internet without much trouble?
Anonymous 01/13/25(Mon)00:17:28 No.103874335
>>103874230
That's the job of the jury. You're absolutely delusional if you think, for example, Mickey Mouse images produced from a lora specifically trained on Mickey Mouse wouldn't violate copyright.
>>103874255
Retard
That's the job of the jury. You're absolutely delusional if you think, for example, Mickey Mouse images produced from a lora specifically trained on Mickey Mouse wouldn't violate copyright.
>>103874255
Retard
Anonymous 01/13/25(Mon)00:18:01 No.103874339
>>103874323
>you cant copyright ai generated content
I think that's good, and that would be the reason AI artists won't die, companies want to copyright their work, so they won't have much choice but to go for real art and no do AI art
>you cant copyright ai generated content
I think that's good, and that would be the reason AI artists won't die, companies want to copyright their work, so they won't have much choice but to go for real art and no do AI art
Anonymous 01/13/25(Mon)00:18:25 No.103874345
>>103874305
right. there will also be many, MANY other countries that will prefer to use generative AI over it all getting outsourced to countries or even paying their own artists.
>>103874323
anon in most countries you automatically have copyright for anything you create with any method, you don't need to apply for it.
idk what kind of a mess the usa will/won't create between states but we'll see
right. there will also be many, MANY other countries that will prefer to use generative AI over it all getting outsourced to countries or even paying their own artists.
>>103874323
anon in most countries you automatically have copyright for anything you create with any method, you don't need to apply for it.
idk what kind of a mess the usa will/won't create between states but we'll see
Anonymous 01/13/25(Mon)00:18:28 No.103874346
>>103874305
learn how machine learning works before posting such crap, I swear we're having the same conversation from 2 years ago
learn how machine learning works before posting such crap, I swear we're having the same conversation from 2 years ago
Anonymous 01/13/25(Mon)00:18:37 No.103874348
>>103874323
You can't copyright AI outputs inherently, but if your AI output is 99% identical to a real copyrighted image, it's obviously a violation of copyright
You can't copyright AI outputs inherently, but if your AI output is 99% identical to a real copyrighted image, it's obviously a violation of copyright
Anonymous 01/13/25(Mon)00:18:51 No.103874351
Did someone let their chatbot loose
Anonymous 01/13/25(Mon)00:19:07 No.103874353
>>103874302
this
this
Anonymous 01/13/25(Mon)00:19:31 No.103874357
>>103874346
that's why I said "perfect" model, nothing can be perfect, but it can be good enough the jury will consider the thing he recreated is close enough to the real deal
that's why I said "perfect" model, nothing can be perfect, but it can be good enough the jury will consider the thing he recreated is close enough to the real deal
Anonymous 01/13/25(Mon)00:20:00 No.103874360
>>103874346
>I swear we're having the same conversation from 2 years ago
seriously. after the first 12 months it gets trite. no one cares anymore
>I swear we're having the same conversation from 2 years ago
seriously. after the first 12 months it gets trite. no one cares anymore
Anonymous 01/13/25(Mon)00:20:26 No.103874365
>>103874323
This is just the copyright office denying them. They're still suing them; lawsuits not over yet.
This is just the copyright office denying them. They're still suing them; lawsuits not over yet.
Anonymous 01/13/25(Mon)00:20:34 No.103874367
>>103874348
Yes but can you prove that in court? That's the difficult part.
Yes but can you prove that in court? That's the difficult part.
Anonymous 01/13/25(Mon)00:20:54 No.103874373
>>103874348
>, but if your AI output is 99% identical to a real copyrighted image, it's obviously a violation of copyright
But is not anon, stop please, learn how ML works specially in diffusion, read about the ongoing lawsuits to stabilityai, openai and midjourney
>, but if your AI output is 99% identical to a real copyrighted image, it's obviously a violation of copyright
But is not anon, stop please, learn how ML works specially in diffusion, read about the ongoing lawsuits to stabilityai, openai and midjourney
Anonymous 01/13/25(Mon)00:22:51 No.103874385
>>103874367
Hence
>>103874335
>That's the job of the jury.
If a jury is willing to accept a lawsuit over two R&B songs sharing similar chords and rhythm, they will absolutely be happy to look at two similar images and rule in favor of the original.
>>103874373
>But is not anon, stop please, learn how ML works specially in diffusion
This has nothing to do with "how it works". If you think picrel wouldn't receive a lawsuit if used on a video game cover, you're retarded.
Hence
>>103874335
>That's the job of the jury.
If a jury is willing to accept a lawsuit over two R&B songs sharing similar chords and rhythm, they will absolutely be happy to look at two similar images and rule in favor of the original.
>>103874373
>But is not anon, stop please, learn how ML works specially in diffusion
This has nothing to do with "how it works". If you think picrel wouldn't receive a lawsuit if used on a video game cover, you're retarded.
Anonymous 01/13/25(Mon)00:26:42 No.103874410
>>103874331
You can do that with anyone's art, regardless of AI. Fuck artists lmao!
You can do that with anyone's art, regardless of AI. Fuck artists lmao!
Anonymous 01/13/25(Mon)00:27:08 No.103874415
>>103874385
what do you mean?
what do you mean?
Anonymous 01/13/25(Mon)00:28:21 No.103874425
Anonymous 01/13/25(Mon)00:29:05 No.103874429
>>103874415
That collage is merely:
bloodborne hunter
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 426535359, Size: 512x512, Model hash: 7460a6fa
No img2img, not even a lora, just raw Stable Diffusion 1.4, and it clearly reproduces a famous copyrighted work.
That collage is merely:
bloodborne hunter
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 426535359, Size: 512x512, Model hash: 7460a6fa
No img2img, not even a lora, just raw Stable Diffusion 1.4, and it clearly reproduces a famous copyrighted work.
Anonymous 01/13/25(Mon)00:29:40 No.103874432
Anonymous 01/13/25(Mon)00:30:38 No.103874445
Anonymous 01/13/25(Mon)00:31:26 No.103874452
>>103874445
That's what my poop looks like.
That's what my poop looks like.
Anonymous 01/13/25(Mon)00:32:28 No.103874458
>>103874173
The real ai power is inpainting anyway. I have a 50mp camera.
The real ai power is inpainting anyway. I have a 50mp camera.
Anonymous 01/13/25(Mon)00:33:39 No.103874470
>>103874432
What does one of the regular noobai do when you do this prompt?
What does one of the regular noobai do when you do this prompt?
Anonymous 01/13/25(Mon)00:34:49 No.103874478
I am two years behind on mastering the art of anime prompting. I spent most of that time making photos of cute girls. At most I've spent maybe a month altogether doing booruprompting with Pony, Easyfluff, NAI, WDXL, Noob, etc.
Can someone who has 2 years' experience mastering the art of anime goon material please give me some of their choicest tips? The ones that took them 1+ year to learn but which can benefit me immediately
Can someone who has 2 years' experience mastering the art of anime goon material please give me some of their choicest tips? The ones that took them 1+ year to learn but which can benefit me immediately
Anonymous 01/13/25(Mon)00:37:22 No.103874497
>>103874470
regular noobai, I can do a cattower one next as well
regular noobai, I can do a cattower one next as well
Anonymous 01/13/25(Mon)00:38:22 No.103874505
>>103874478
Anime has become so simple that it's the default thing to spam
Anime has become so simple that it's the default thing to spam
Anonymous 01/13/25(Mon)00:38:54 No.103874508
Anonymous 01/13/25(Mon)00:39:50 No.103874521
>>103874497
looks neat. But not as varied in color.
looks neat. But not as varied in color.
Anonymous 01/13/25(Mon)00:41:30 No.103874534
>>103874130
will grok get sued? everyone keeps making celeb stuff with grok on twitter and deviant
will grok get sued? everyone keeps making celeb stuff with grok on twitter and deviant
Anonymous 01/13/25(Mon)00:42:18 No.103874539
>>103874478
>Can someone who has 2 years' experience mastering the art of anime goon material please give me some of their choicest tips?
learn to inpaint, upscaling with a tiled cn is alot more stable than normal hires fix so this allows you to do all your inpainting at base resolution, this helps keep the lighting more consitant when doing major changes to your image
>Can someone who has 2 years' experience mastering the art of anime goon material please give me some of their choicest tips?
learn to inpaint, upscaling with a tiled cn is alot more stable than normal hires fix so this allows you to do all your inpainting at base resolution, this helps keep the lighting more consitant when doing major changes to your image
Anonymous 01/13/25(Mon)00:42:20 No.103874543
Anonymous 01/13/25(Mon)00:42:23 No.103874545
>>103874534
I don't think Elon gives a fuck, he's the richest man ever, he can handle lawsuits
I don't think Elon gives a fuck, he's the richest man ever, he can handle lawsuits
Anonymous 01/13/25(Mon)00:42:34 No.103874546
>>103874429
Nah. You're too stupid to even understand.
Nah. You're too stupid to even understand.
Anonymous 01/13/25(Mon)00:44:41 No.103874570
/ldg/ - Ran (the biggest avatar known to 4chan) and some political posters who have trouble genning a coherent images.
Anonymous 01/13/25(Mon)00:44:42 No.103874571
>>103874546
Pretty sure it's some sperg from /ic/
Pretty sure it's some sperg from /ic/
Anonymous 01/13/25(Mon)00:44:55 No.103874574
Anonymous 01/13/25(Mon)00:44:58 No.103874577
>>103874385
>>103874429
anon, the only retarded person in this thread is you, clearly you don't know what are you talking about and you're mixing concepts
>>103874429
anon, the only retarded person in this thread is you, clearly you don't know what are you talking about and you're mixing concepts
Anonymous 01/13/25(Mon)00:45:16 No.103874580
>>103874545
he can handle it but I wonder if celebs are trying to sue
he can handle it but I wonder if celebs are trying to sue
Anonymous 01/13/25(Mon)00:47:04 No.103874593
>>103874331
the usa made this a messy "maybe", for now it seems like any human effort in the work may make it copyrightable else not. they really didn't define it well and bullshit until they can maybe pass laws that please them, but the big corpos who produce AI images will likely also influence this decision with all the pull they have in US politics.
china and others say it's just copyrighted period... now you may say "it's china, what do I care?" but they're a berne convention member, and in the berne convention it only falls on other countries to protect stuff copyrighted artwork in other member nations, not investigate the methods by which they were produced or anything like that...
the usa made this a messy "maybe", for now it seems like any human effort in the work may make it copyrightable else not. they really didn't define it well and bullshit until they can maybe pass laws that please them, but the big corpos who produce AI images will likely also influence this decision with all the pull they have in US politics.
china and others say it's just copyrighted period... now you may say "it's china, what do I care?" but they're a berne convention member, and in the berne convention it only falls on other countries to protect stuff copyrighted artwork in other member nations, not investigate the methods by which they were produced or anything like that...
Anonymous 01/13/25(Mon)00:48:23 No.103874610
>>103874593
It's china, why do I care?
nah I'm kidding, that was interesting to read, I learned something today so thanks for that
It's china, why do I care?
nah I'm kidding, that was interesting to read, I learned something today so thanks for that
Anonymous 01/13/25(Mon)00:52:37 No.103874639
>>103874580
sue for what exactly? Whats up with all these anons talking about lawsuits like is some shit you buy in a store, you guys watch too much movies and tv series
sue for what exactly? Whats up with all these anons talking about lawsuits like is some shit you buy in a store, you guys watch too much movies and tv series
Anonymous 01/13/25(Mon)00:54:28 No.103874654
anyone thinking of scraping social media for random women and targeting them and their families with blackmail via ai images?
Anonymous 01/13/25(Mon)00:55:22 No.103874658
Anonymous 01/13/25(Mon)00:55:46 No.103874661
>>103874654
Thanks kinda psychotic desu
Thanks kinda psychotic desu
Anonymous 01/13/25(Mon)00:55:56 No.103874662
Anonymous 01/13/25(Mon)00:56:06 No.103874663
>>103874639
>sue for what exactly?
https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html
>sue for what exactly?
https://www.nytimes.com/2023/12/27/
Anonymous 01/13/25(Mon)00:56:40 No.103874667
We can do it boys!
Anonymous 01/13/25(Mon)00:56:46 No.103874668
>>103874661
That's*
That's*
Anonymous 01/13/25(Mon)00:58:20 No.103874679
>>103874546
>>103874577
>you don't know what you're talking about
>no I can't explain it but trust me you don't
>>103874577
>you don't know what you're talking about
>no I can't explain it but trust me you don't
Anonymous 01/13/25(Mon)00:58:21 No.103874680
Anonymous 01/13/25(Mon)00:58:30 No.103874681
Anonymous 01/13/25(Mon)00:58:32 No.103874682
>>103874663
so?
https://www.hollywoodreporter.com/business/business-news/sarah-silverman-lawsuit-openai-copyrighted-novels-1235963290/
so?
https://www.hollywoodreporter.com/b
Anonymous 01/13/25(Mon)00:58:43 No.103874684
>>103874521
played around a little bit, and you can get some with decent color in NAI if you specify "lots of color" and "neon colors" or anything to add pops of color. i still think blackmagic has the better overall design though
played around a little bit, and you can get some with decent color in NAI if you specify "lots of color" and "neon colors" or anything to add pops of color. i still think blackmagic has the better overall design though
Anonymous 01/13/25(Mon)01:00:01 No.103874695
>>103874682
yeah, they won't win those lawsuits, but the goal is to make someone poor by forcing them to pay for expensive lawyers, OpenAI is fine they are rich, but the random fags who makes celebrities lora will bend the knee if a celebrity tell him they'll sue him and he'll have to pay hundreds of thousands of dollars to win the case
yeah, they won't win those lawsuits, but the goal is to make someone poor by forcing them to pay for expensive lawyers, OpenAI is fine they are rich, but the random fags who makes celebrities lora will bend the knee if a celebrity tell him they'll sue him and he'll have to pay hundreds of thousands of dollars to win the case
Anonymous 01/13/25(Mon)01:03:08 No.103874716
>>103874695
sue who exactly? some random guy from an eastern european country, better yet what if the person who made the lora is from china? who and how are they gonna sue said person, are they going to extradite him? lmao
sue who exactly? some random guy from an eastern european country, better yet what if the person who made the lora is from china? who and how are they gonna sue said person, are they going to extradite him? lmao
Anonymous 01/13/25(Mon)01:03:34 No.103874721
Anonymous 01/13/25(Mon)01:04:15 No.103874724
>>103874716
>who and how are they gonna sue said person
you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
>who and how are they gonna sue said person
you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
Anonymous 01/13/25(Mon)01:06:52 No.103874739
>>103874724
>you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
AHAHAHAHAHAHAHH anon, you clearly live a in a bubble lmao
>you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
AHAHAHAHAHAHAHH anon, you clearly live a in a bubble lmao
Anonymous 01/13/25(Mon)01:07:59 No.103874746
Im starting to like flux.
Anonymous 01/13/25(Mon)01:08:15 No.103874750
Anonymous 01/13/25(Mon)01:09:09 No.103874761
>>103874739
Hiroshi Fujiwara v. Vivian Westwood (Japan vs. UK)
Barbara Streisand v. Kenneth Adelman (USA vs. Sweden)
Erin Gallagher v. Patrick Murphy (Ireland vs. USA)
Judith Krantz v. Jacob Dahmer (USA vs. Germany)
Laleh Shahravesh v. Samah Al Hammadi (UK vs. UAE)
Gina Lollobrigida v. Javier Rigau (Italy vs. Spain)
Hiroshi Fujiwara v. Vivian Westwood (Japan vs. UK)
Barbara Streisand v. Kenneth Adelman (USA vs. Sweden)
Erin Gallagher v. Patrick Murphy (Ireland vs. USA)
Judith Krantz v. Jacob Dahmer (USA vs. Germany)
Laleh Shahravesh v. Samah Al Hammadi (UK vs. UAE)
Gina Lollobrigida v. Javier Rigau (Italy vs. Spain)
Anonymous 01/13/25(Mon)01:13:48 No.103874792
>>103874746
it is powerful but the de-distilled models need a lot more 1girl / 1boy finetuning and this seems difficult to do
it is powerful but the de-distilled models need a lot more 1girl / 1boy finetuning and this seems difficult to do
Anonymous 01/13/25(Mon)01:14:14 No.103874798
>>103874761
You had ChatGPT generate that
You had ChatGPT generate that
Anonymous 01/13/25(Mon)01:14:54 No.103874805
>>103874798
Moving the goalpost I see
Moving the goalpost I see
Anonymous 01/13/25(Mon)01:15:03 No.103874807
>>103874684
Neat!
Neat!
Anonymous 01/13/25(Mon)01:15:24 No.103874810
>>103874805
Link me an article for the first one
Link me an article for the first one
Anonymous 01/13/25(Mon)01:18:01 No.103874826
>>103874810
you said it's impossible to sue someone from another country, just one counterexample is enough to deboonk your retarded assumption, are you done anon?
you said it's impossible to sue someone from another country, just one counterexample is enough to deboonk your retarded assumption, are you done anon?
Anonymous 01/13/25(Mon)01:18:05 No.103874828
>>103874798
Oh that one flew over my head kek
Oh that one flew over my head kek
Anonymous 01/13/25(Mon)01:19:12 No.103874840
>>103874826
>can't give a link
The docket would work too, I find nothing. You're full of shit
>you
NTA btw
>can't give a link
The docket would work too, I find nothing. You're full of shit
>you
NTA btw
Anonymous 01/13/25(Mon)01:21:17 No.103874855
>>103874840
google is free anon
https://www.pinsentmasons.com/out-law/news/internet-publishers-liable-for-privacy-invasion-in-each-country-material-is-accessible-ecj-rules?utm_source=chatgpt.com
>French actor Oliver Martinez who had sued the UK's Sunday Mirror in a French court over a story about his relationship with Australian singer Kylie Minogue.
google is free anon
https://www.pinsentmasons.com/out-l
>French actor Oliver Martinez who had sued the UK's Sunday Mirror in a French court over a story about his relationship with Australian singer Kylie Minogue.
Anonymous 01/13/25(Mon)01:22:15 No.103874865
>>103874761
wow anon a total of 10 countries, where are the cases against china or russia? which was my original question
>>103874798
kek, thought the same thing, next time be more specific to chatgpt please, tell it to be about copyright which is the subject we're talking about
wow anon a total of 10 countries, where are the cases against china or russia? which was my original question
>>103874798
kek, thought the same thing, next time be more specific to chatgpt please, tell it to be about copyright which is the subject we're talking about
Anonymous 01/13/25(Mon)01:23:14 No.103874877
>>103874865
>wow anon a total of 10 countries,
so we went from "iz impossibruuu" to "you only linked 10 cases, why didn't you link all the cases in the history of the humanity?", you're funny not gonna lie
>wow anon a total of 10 countries,
so we went from "iz impossibruuu" to "you only linked 10 cases, why didn't you link all the cases in the history of the humanity?", you're funny not gonna lie
Anonymous 01/13/25(Mon)01:24:02 No.103874885
Anonymous 01/13/25(Mon)01:24:29 No.103874887
>>103874885
Concession Accepted.
Concession Accepted.
Anonymous 01/13/25(Mon)01:25:30 No.103874893
>>103874885
I accept your concession anon
https://www.theguardian.com/media/2011/oct/25/sunday-mirror-olivier-martinez#:~:text=The%20publisher%20of%20the%20Sunday,the%20pop%20star%20Kylie%20Minogue.
>The publisher of the Sunday Mirror has lost its bid to overturn a privacy action brought in the French courts by the actor Olivier Martinez.
I accept your concession anon
https://www.theguardian.com/media/2
>The publisher of the Sunday Mirror has lost its bid to overturn a privacy action brought in the French courts by the actor Olivier Martinez.
Anonymous 01/13/25(Mon)01:25:58 No.103874897
>>103874877
>you said it's impossible to sue someone from another country,
>so we went from "iz impossibruuu"
Can you pinpoint the post where I exactly said it was "impossible" ? the only one who said that was yourself
>you said it's impossible to sue someone from another country,
>so we went from "iz impossibruuu"
Can you pinpoint the post where I exactly said it was "impossible" ? the only one who said that was yourself
Anonymous 01/13/25(Mon)01:26:42 No.103874905
>>103874897
>Can you pinpoint the post where I exactly said it was "impossible" ?
sure
>>103874739
>>you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
>AHAHAHAHAHAHAHH anon, you clearly live a in a bubble lmao
>Can you pinpoint the post where I exactly said it was "impossible" ?
sure
>>103874739
>>you can't sue someone who isn't on the same country as yours? what kind of nonsense is that?
>AHAHAHAHAHAHAHH anon, you clearly live a in a bubble lmao
Anonymous 01/13/25(Mon)01:26:47 No.103874906
>>103874893
Good job taking chatgpt out of the URL this time anon
>>103874897
He's just wasting time at this point
Good job taking chatgpt out of the URL this time anon
>>103874897
He's just wasting time at this point
Anonymous 01/13/25(Mon)01:31:06 No.103874930
>>103874905
kek, I fail to see where I posted "It was impossible", I know you're baiting now, but poor anon, you need to go back to school, clearly you don't understand what you read, whats worst is that you make assumptions in your head from something that isn't even there
kek, I fail to see where I posted "It was impossible", I know you're baiting now, but poor anon, you need to go back to school, clearly you don't understand what you read, whats worst is that you make assumptions in your head from something that isn't even there
Anonymous 01/13/25(Mon)01:33:14 No.103874951
>>103874930
>I fail to see where I posted "It was impossible"
so when someone asks you if it's possible to sue someone in another country, to tell him it's possible you laugh at his face? I've seen some weirds anons on 4chan, but you're kinda special not gonna lie
>I fail to see where I posted "It was impossible"
so when someone asks you if it's possible to sue someone in another country, to tell him it's possible you laugh at his face? I've seen some weirds anons on 4chan, but you're kinda special not gonna lie
Anonymous 01/13/25(Mon)01:35:14 No.103874963
>>103873497
God bless you.
God bless you.
Anonymous 01/13/25(Mon)01:36:42 No.103874979
It's a work in progress
the jeets who work here actively try to ban people, because they have perverse incentives: fewer people, easier job. So they look for excuses to ban people, which is very funny, but it explains why so few are here now. jeets suck
the jeets who work here actively try to ban people, because they have perverse incentives: fewer people, easier job. So they look for excuses to ban people, which is very funny, but it explains why so few are here now. jeets suck
Anonymous 01/13/25(Mon)01:37:43 No.103874987
>>103874963
That guy looks familiar. I wonder if he knows hwnbaw.
That guy looks familiar. I wonder if he knows hwnbaw.
Anonymous 01/13/25(Mon)01:38:12 No.103874994
What models are used for inpainting these days?
Anonymous 01/13/25(Mon)01:41:23 No.103875011
Anonymous 01/13/25(Mon)01:42:03 No.103875014
>>103874979
Why are you obsessed with Indians? Isn't this quite absurd considering your own location and all that.
Why are you obsessed with Indians? Isn't this quite absurd considering your own location and all that.
Anonymous 01/13/25(Mon)01:44:32 No.103875027
Illustrious/Noob is just Pony/Autismmix all over again. Noob is obviously inferior, but maybe easier to """get good results""", Illustrious far more fun to use and capable of more variety.
Anonymous 01/13/25(Mon)01:45:25 No.103875036
Anonymous 01/13/25(Mon)01:46:29 No.103875042
>indians are using ai to stalk 4chan
fresh jeet hell
fresh jeet hell
Anonymous 01/13/25(Mon)01:58:59 No.103875133
Anonymous 01/13/25(Mon)02:01:50 No.103875166
So someone asked this in /pcbg/
8GB RTX 3070 Ti or 16GB RX 6800?
8GB RTX 3070 Ti or 16GB RX 6800?
Anonymous 01/13/25(Mon)02:03:49 No.103875178
Anonymous 01/13/25(Mon)02:05:18 No.103875192
>>103875178
Are AMD APIs and ZLUDA really that shit for this?
Are AMD APIs and ZLUDA really that shit for this?
Anonymous 01/13/25(Mon)02:05:39 No.103875197
>>103875027
That's a cute opinion.
That's a cute opinion.
Anonymous 01/13/25(Mon)02:05:39 No.103875199
>>103875192
Yes
Yes
Anonymous 01/13/25(Mon)02:06:51 No.103875212
>>103875192
no, it's because nobody learns rocm in school, and once they leave they never learn anything else ever, they are cuda forever.
no, it's because nobody learns rocm in school, and once they leave they never learn anything else ever, they are cuda forever.
Anonymous 01/13/25(Mon)02:08:47 No.103875224
>>103875197
It's also a true one.
It's also a true one.
Anonymous 01/13/25(Mon)02:08:51 No.103875226
>>103875212
I meant it regarding software support and being able to use what's already out, not "can a programmer create new software that uses ROCm instead?"
I meant it regarding software support and being able to use what's already out, not "can a programmer create new software that uses ROCm instead?"
Anonymous 01/13/25(Mon)02:09:47 No.103875236
>>103875226
I answered both.
I answered both.
Anonymous 01/13/25(Mon)02:11:24 No.103875254
>>103875224
You are so stupid that you are starting to argue with your own posts? I don't know are you all underage dimwits or what is your issue but I'm actually happy that /ldg/ attracts all underage pol rejects like you.
You are so stupid that you are starting to argue with your own posts? I don't know are you all underage dimwits or what is your issue but I'm actually happy that /ldg/ attracts all underage pol rejects like you.
Anonymous 01/13/25(Mon)02:15:52 No.103875289
>>103875254
>I'm actually happy that /ldg/ attracts all underage pol rejects like you.
why? you're a /sdg/ fag who wants to see /ldg/'s downfall?
>I'm actually happy that /ldg/ attracts all underage pol rejects like you.
why? you're a /sdg/ fag who wants to see /ldg/'s downfall?
Hi My name IS mUD 01/13/25(Mon)02:22:01 No.103875336
RocmSocmRobots
Anonymous 01/13/25(Mon)02:26:21 No.103875366
>>103875289
Still too cretinous to understand. There is no downfall - this is the bottom of the barrel because of you.
Still too cretinous to understand. There is no downfall - this is the bottom of the barrel because of you.
Anonymous 01/13/25(Mon)02:28:06 No.103875380
Anonymous 01/13/25(Mon)02:30:05 No.103875396
Anonymous 01/13/25(Mon)02:31:25 No.103875400
>>103875396
kek
kek
Anonymous 01/13/25(Mon)02:38:16 No.103875447
>>103875254
Are you quoting the wrong post? What are you talking about?
Are you quoting the wrong post? What are you talking about?
Anonymous 01/13/25(Mon)02:44:51 No.103875478
I'm remembering why I don't gen anime 1girl so much, it's too easy to get distracted with porn and lose interest. Keeping it postable 1girl is harder. On flux/etc. sfw solo 1girl is all you can gen anyway.
Anonymous 01/13/25(Mon)02:49:43 No.103875514
>>103875447
Are you talking to me? I don't see anyone else here so you must be talking to me.
Are you talking to me? I don't see anyone else here so you must be talking to me.
Anonymous 01/13/25(Mon)02:51:24 No.103875524
>>103875478
kino style
kino style
Anonymous 01/13/25(Mon)02:57:01 No.103875550
>>103875478
I don't get because there's no point in genning it. It's all slop and there's nothing interesting about it. Anime 1girl does not make me horny, and I say this as a guy who used to fap to hentai. Just doesn't have the same effect as when a human draws it, the AI is just not capable of capturing it. That being said, Dalle was pretty good. I guess it has to do with SD/Flux being censored from scratch so them not being able to naturally draw vaginas, they actually suck at making porn, if they make porn it's always in a shitty uninspired style. Whereas Dalle has seen the thing so it knows how to do it the same way a human would pretty well.
I don't get because there's no point in genning it. It's all slop and there's nothing interesting about it. Anime 1girl does not make me horny, and I say this as a guy who used to fap to hentai. Just doesn't have the same effect as when a human draws it, the AI is just not capable of capturing it. That being said, Dalle was pretty good. I guess it has to do with SD/Flux being censored from scratch so them not being able to naturally draw vaginas, they actually suck at making porn, if they make porn it's always in a shitty uninspired style. Whereas Dalle has seen the thing so it knows how to do it the same way a human would pretty well.
Anonymous 01/13/25(Mon)02:59:07 No.103875558
Anonymous 01/13/25(Mon)03:04:30 No.103875585
>>103875550
What kind of a mongoloid even bothers to type all this?
What kind of a mongoloid even bothers to type all this?
Anonymous 01/13/25(Mon)03:05:33 No.103875589
>>103875585
An uninspired consoomer
An uninspired consoomer
Anonymous 01/13/25(Mon)03:10:01 No.103875623
>>103875585
>>103875589
It's true, and you know it. Even clothed 1girls posted by shitty censored open models do not stand a chance against Dalle.
>>103875589
It's true, and you know it. Even clothed 1girls posted by shitty censored open models do not stand a chance against Dalle.
Anonymous 01/13/25(Mon)03:14:58 No.103875652
Ran has ruined this general.
Anonymous 01/13/25(Mon)03:31:38 No.103875748
>>103875623
skill issue
skill issue
Anonymous 01/13/25(Mon)03:38:11 No.103875785
the lawsuit is about porn? but then people on aco are still using dall e to create porn using a website called cici ai and bot on coze. but how do they do that
Anonymous 01/13/25(Mon)03:38:18 No.103875786
Anonymous 01/13/25(Mon)03:41:31 No.103875808
>>103875166
Serious answer. With RDNA 3 and up, it should be good enough for inferencing with Flash Attention support and other stuff but with the 6800, the answer is no. Even then, you have to ask yourself if the savings worth waiting for software support with Nvidia 1st and ZLUDA/translation issues? Only you can decide that but for most people, the answer to that question is no.
Serious answer. With RDNA 3 and up, it should be good enough for inferencing with Flash Attention support and other stuff but with the 6800, the answer is no. Even then, you have to ask yourself if the savings worth waiting for software support with Nvidia 1st and ZLUDA/translation issues? Only you can decide that but for most people, the answer to that question is no.
Anonymous 01/13/25(Mon)03:46:41 No.103875841
>>103875166
3060 12GB
3060 12GB
Anonymous 01/13/25(Mon)04:18:39 No.103876000
Anonymous 01/13/25(Mon)04:21:41 No.103876016
>>103874963
y-you t-too!
y-you t-too!
Anonymous 01/13/25(Mon)05:24:33 No.103876405
Anonymous 01/13/25(Mon)05:28:51 No.103876425
Anonymous 01/13/25(Mon)05:39:48 No.103876497
>>103874750
migu no :(
migu no :(
Anonymous 01/13/25(Mon)05:46:06 No.103876545
>>103876016
all her parents effort down the drain
all her parents effort down the drain
Anonymous 01/13/25(Mon)05:50:03 No.103876580
I haven’t used colab in years. I’m assuming I can just inpaint magic poser screenshots with good protmths for scenes?
Anonymous 01/13/25(Mon)05:51:30 No.103876590
>>103876425
anon... take her to the coom farm
anon... take her to the coom farm
Anonymous 01/13/25(Mon)06:02:21 No.103876645
>>103876425
what's the prompt here or the data
what's the prompt here or the data
Anonymous 01/13/25(Mon)06:02:50 No.103876647
Anonymous 01/13/25(Mon)06:16:32 No.103876715
>>103876645
Made that one in Krita with an AI plugin, sadly it doesn't save anything in the metadata. The prompt was
"masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest
(perseus \(azur lane\):1.2), alternate outfit
(traditional media, pinup \(style\), 1940s, 1950s, art by norman rockwell, watercolor \(medium\), painterly, highly detailed, smooth, impressionism, vibrant, high contrast, sketch:1.2)
(artist:nyantcha, artist:shexyo, artist:dishwasher1910:0.9)
1girl, posing, on bed, close-up, head tilt, sitting, from side, adjusting footwear
(black background, pink background:1.1), abstract background
looking at viewer, half-closed eyes, parted lips, puckered lips, pink eyes
(pink lips:1.2), shiny lips, thick lips, lipstick, long eyelashes, black eyeliner, mascara, eyeshadow, highly detailed eyes
(pink lingerie:1.2), pink corset, elbow gloves, fingerless gloves, pink gloves, bra, panties, stockings, pink thighhighs, (decorative bow:1.2), gothic lolita, lace, lace choker, garter strap
skindentation, shiny skin
pink nails, long nails, shiny nails, nail art
pink hair, absurdly long hair, blunt bangs, twintails
large breasts, curvy, thick thighs
masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest"
and the model used was NTRMix v4.0
Made that one in Krita with an AI plugin, sadly it doesn't save anything in the metadata. The prompt was
"masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest
(perseus \(azur lane\):1.2), alternate outfit
(traditional media, pinup \(style\), 1940s, 1950s, art by norman rockwell, watercolor \(medium\), painterly, highly detailed, smooth, impressionism, vibrant, high contrast, sketch:1.2)
(artist:nyantcha, artist:shexyo, artist:dishwasher1910:0.9)
1girl, posing, on bed, close-up, head tilt, sitting, from side, adjusting footwear
(black background, pink background:1.1), abstract background
looking at viewer, half-closed eyes, parted lips, puckered lips, pink eyes
(pink lips:1.2), shiny lips, thick lips, lipstick, long eyelashes, black eyeliner, mascara, eyeshadow, highly detailed eyes
(pink lingerie:1.2), pink corset, elbow gloves, fingerless gloves, pink gloves, bra, panties, stockings, pink thighhighs, (decorative bow:1.2), gothic lolita, lace, lace choker, garter strap
skindentation, shiny skin
pink nails, long nails, shiny nails, nail art
pink hair, absurdly long hair, blunt bangs, twintails
large breasts, curvy, thick thighs
masterpiece, best quality, amazing quality, very aesthetic, absurdres, newest"
and the model used was NTRMix v4.0
Anonymous 01/13/25(Mon)06:58:08 No.103876958
Anonymous 01/13/25(Mon)08:18:49 No.103877559
>>103873827
You will have people that are actually smart in your country.
You will have people that are actually smart in your country.
Anonymous 01/13/25(Mon)08:20:33 No.103877574
>>103877559
so Elon and Trump are smart? they are multi billionaires
so Elon and Trump are smart? they are multi billionaires
Anonymous 01/13/25(Mon)08:32:25 No.103877642
>>103877574
One managed to become president twice and only the second one to manage it non-consecutively and the other is literally the richest person on planet earth.
Like what are you trying to say? They just blundered into those positions? You can disagree with their politics, but to call them not smart is discard every metric of success and replace it with your own arbitrary yardstick that fits into your narrow definition of intelligent.
Or is this just like a ludicrous post where you bait people to response to an insane claim?
One managed to become president twice and only the second one to manage it non-consecutively and the other is literally the richest person on planet earth.
Like what are you trying to say? They just blundered into those positions? You can disagree with their politics, but to call them not smart is discard every metric of success and replace it with your own arbitrary yardstick that fits into your narrow definition of intelligent.
Or is this just like a ludicrous post where you bait people to response to an insane claim?
Anonymous 01/13/25(Mon)08:33:33 No.103877654
>>103877642
I just asked a question, I voted for Trump lol, you didn't need to be so defensive
I just asked a question, I voted for Trump lol, you didn't need to be so defensive
Anonymous 01/13/25(Mon)08:35:29 No.103877667
>>103877642
>They just blundered into those positions?
Trump definitely did get rich thanks to his parents yeah
https://www.youtube.com/watch?v=D3wY8cuB-lQ
>They just blundered into those positions?
Trump definitely did get rich thanks to his parents yeah
https://www.youtube.com/watch?v=D3w
Anonymous 01/13/25(Mon)08:38:11 No.103877687
>>103877654
It's just such a common thing people say despite objective indicating they are at least somewhat smart.
>>103877667
If you can turn one million into billions you're probably smart.
It's just such a common thing people say despite objective indicating they are at least somewhat smart.
>>103877667
If you can turn one million into billions you're probably smart.
Anonymous 01/13/25(Mon)08:39:24 No.103877697
>>103877687
>If you can turn one million into billions you're probably smart.
I disagree with that, it's way easier to invest when you have a shit ton of money at the begining of the process
>If you can turn one million into billions you're probably smart.
I disagree with that, it's way easier to invest when you have a shit ton of money at the begining of the process
Anonymous 01/13/25(Mon)08:44:40 No.103877734
>noobshit doesn't know the hat on Hasshaku-sama
No wonder it's called as such, only a newfag wouldn't know this character.
No wonder it's called as such, only a newfag wouldn't know this character.
Anonymous 01/13/25(Mon)08:50:11 No.103877773
>>103877697
>easier
Not easy, easier.
Look what happens to most people who come into large amounts of money in the first few years.
>easier
Not easy, easier.
Look what happens to most people who come into large amounts of money in the first few years.
Anonymous 01/13/25(Mon)08:54:05 No.103877806
>>103877773
how hard is it to fail when you get free 1 millions dollars and your parents know how to invest and will teach you the ways? I like Trump but he's far from a self made man, Elon on the other hand...
how hard is it to fail when you get free 1 millions dollars and your parents know how to invest and will teach you the ways? I like Trump but he's far from a self made man, Elon on the other hand...
Anonymous 01/13/25(Mon)08:56:09 No.103877827
>>103877806
I feel like you're trying to lead me down a trap here. Both of them made good investments with their money that paid off much higher in return than you might otherwise expect. They're smart.
I feel like you're trying to lead me down a trap here. Both of them made good investments with their money that paid off much higher in return than you might otherwise expect. They're smart.
Anonymous 01/13/25(Mon)08:58:40 No.103877851
>>103877806
elon's dad was rich af though, but he was indeed self made, unlike elon
>source: wikipedia
elon's dad was rich af though, but he was indeed self made, unlike elon
>source: wikipedia
Anonymous 01/13/25(Mon)08:59:20 No.103877859
>>103877827
Idk man, intelligence isn't just about being a good investor, life is much more complex than that, Grigori Perelman for example is the only guy that solved one of the 7 Millennium Prize Problems, yet he's poor as fuck
Idk man, intelligence isn't just about being a good investor, life is much more complex than that, Grigori Perelman for example is the only guy that solved one of the 7 Millennium Prize Problems, yet he's poor as fuck
Anonymous 01/13/25(Mon)08:59:36 No.103877862
i'm sorry for disussing image generation models in a /pol/ branch
Anonymous 01/13/25(Mon)09:04:31 No.103877904
i have all images filtered out
Anonymous 01/13/25(Mon)09:05:26 No.103877915
>>103877904
based
based
Anonymous 01/13/25(Mon)09:15:11 No.103877999
>>103876715
my 1girl isn't as pretty as your 1girl...
my 1girl isn't as pretty as your 1girl...
Anonymous 01/13/25(Mon)09:18:42 No.103878031
>>103877862
>/pol/ tier spamming
>only image posters are Ran (which ironically is the biggest and most annoying avatarfag of them all) and couple of others with weird fixation on semen
This is the irony of it all.
>/pol/ tier spamming
>only image posters are Ran (which ironically is the biggest and most annoying avatarfag of them all) and couple of others with weird fixation on semen
This is the irony of it all.
Anonymous 01/13/25(Mon)09:30:51 No.103878143
Is this the correct way to prompt?
>masterpiece, best quality, 1girl,
>sugimori ken,
>oleana,
>black footwear, black pantyhose, black skirt, blonde hair, breasts, closed mouth, coat, collarbone, crystal, earrings, female focus, forehead, full body, green eyes, half-closed eyes, hand on own hip, hoop earrings, jewelry, jitome, legs apart, lipstick, long hair, looking at viewer, makeup, medium breasts, miniskirt, nail polish, pantyhose, parted bangs, pencil skirt, red lips, red nails, red shirt, shirt, shoes, short sleeves, side slit, skirt, solo, standing, white coat
First line should be quality tags, then a line break, then the artist name, then a line break, then the name of the character, then a line break, then all the tags that describe the specifics of the image?
>masterpiece, best quality, 1girl,
>sugimori ken,
>oleana,
>black footwear, black pantyhose, black skirt, blonde hair, breasts, closed mouth, coat, collarbone, crystal, earrings, female focus, forehead, full body, green eyes, half-closed eyes, hand on own hip, hoop earrings, jewelry, jitome, legs apart, lipstick, long hair, looking at viewer, makeup, medium breasts, miniskirt, nail polish, pantyhose, parted bangs, pencil skirt, red lips, red nails, red shirt, shirt, shoes, short sleeves, side slit, skirt, solo, standing, white coat
First line should be quality tags, then a line break, then the artist name, then a line break, then the name of the character, then a line break, then all the tags that describe the specifics of the image?
Anonymous 01/13/25(Mon)09:33:37 No.103878178
>>103878143
Do you have to type
BREAK
between line breaks in order to make it "count"?
What's the benefit of doing it/downside of not doing it?
Do you have to type
BREAK
between line breaks in order to make it "count"?
What's the benefit of doing it/downside of not doing it?
Anonymous 01/13/25(Mon)09:41:15 No.103878255
>>103878178
Manually formatting your prompt with spaces is useful for you as that will make things easier to organize.
BREAK doesn't work in ComfyUI as every token is averaged. A1111 and its successors are using a different weighting method and BREAK resets the current weights.
If you want similar approach in ComfyUI you can use conditioning concat node.
But you should pretty much manage with basic prompting and manual weights here and there, no need for additional shenanigans. It can be helpful with multiple subjects though.
Manually formatting your prompt with spaces is useful for you as that will make things easier to organize.
BREAK doesn't work in ComfyUI as every token is averaged. A1111 and its successors are using a different weighting method and BREAK resets the current weights.
If you want similar approach in ComfyUI you can use conditioning concat node.
But you should pretty much manage with basic prompting and manual weights here and there, no need for additional shenanigans. It can be helpful with multiple subjects though.
Anonymous 01/13/25(Mon)09:41:22 No.103878256
>>103877999
I did a lot of manual inpainting after the initial generation (pic rel) to fix up details etc. What generation settings (CFG etc) are you using? And do you generate at that resolution or do you upscale images automatically?
I did a lot of manual inpainting after the initial generation (pic rel) to fix up details etc. What generation settings (CFG etc) are you using? And do you generate at that resolution or do you upscale images automatically?
Anonymous 01/13/25(Mon)09:47:24 No.103878298
>>103877734
Where's the catbox shill, now that the image is posted right away he's not able to pull his only card anymore?
Noobai is so done.
Where's the catbox shill, now that the image is posted right away he's not able to pull his only card anymore?
Noobai is so done.
Anonymous 01/13/25(Mon)09:48:47 No.103878309
>>103878143
You know the model is shit if it requires ANY prompt structure. The caption shuffling during training is made to avoid that need.
You know the model is shit if it requires ANY prompt structure. The caption shuffling during training is made to avoid that need.
Anonymous 01/13/25(Mon)09:54:51 No.103878368
>>103878256
I usually use a CFG scale of 7. Is that a bad practice? I am never certain what is better, going higher or lower. I'm not sure where the sweet spot is.
I tend to gen at resolutions like 832x1216 and then upscale 2x.
I usually use a CFG scale of 7. Is that a bad practice? I am never certain what is better, going higher or lower. I'm not sure where the sweet spot is.
I tend to gen at resolutions like 832x1216 and then upscale 2x.
Anonymous 01/13/25(Mon)09:55:44 No.103878376
>>103878368
This image and the above image were both upscaled from the same base image...the only difference was denoising strength
crazy how different the final image can turn out if you pump up the denoising amount
This image and the above image were both upscaled from the same base image...the only difference was denoising strength
crazy how different the final image can turn out if you pump up the denoising amount
Anonymous 01/13/25(Mon)09:56:22 No.103878381
Anonymous 01/13/25(Mon)09:56:41 No.103878384
i'm going to slop and you can't stop me
Anonymous 01/13/25(Mon)09:57:23 No.103878392
>>103878256
nta, very interesting to see pre-process gens
nta, very interesting to see pre-process gens
Anonymous 01/13/25(Mon)10:20:43 No.103878574
>>103869129
Okay your insistence made me revisit the model and consequentially reevaluate my views.
Okay your insistence made me revisit the model and consequentially reevaluate my views.
Anonymous 01/13/25(Mon)10:27:07 No.103878639
I am eagerly waiting for Catjak's latest masterpiece.
Anonymous 01/13/25(Mon)10:27:37 No.103878647
>>103878368
Nah, it's not a bad practice whatsoever, it's just a matter of different model and user preferences. NTRMix specifically seems to work well with a CFG scale of around 4. If you've got it set to 7, that would explain why the contrast is so much stronger. I think most modern models work well with lower values in general, maybe because they've got better prompt understanding overall. When trying out a new model, I usually just browse the example pictures on civitai and steal the generation settings from there.
>I tend to gen at resolutions like 832x1216 and then upscale 2x.
Ah, ok. Do you do the detailing before or after the upscaling? Upscaling tends to mess up the details in my experience, so I generally don't bother?
Nah, it's not a bad practice whatsoever, it's just a matter of different model and user preferences. NTRMix specifically seems to work well with a CFG scale of around 4. If you've got it set to 7, that would explain why the contrast is so much stronger. I think most modern models work well with lower values in general, maybe because they've got better prompt understanding overall. When trying out a new model, I usually just browse the example pictures on civitai and steal the generation settings from there.
>I tend to gen at resolutions like 832x1216 and then upscale 2x.
Ah, ok. Do you do the detailing before or after the upscaling? Upscaling tends to mess up the details in my experience, so I generally don't bother?
Anonymous 01/13/25(Mon)10:29:48 No.103878669
>>103878639
certainly worth being included in the collage I'll bet
certainly worth being included in the collage I'll bet
Anonymous 01/13/25(Mon)10:29:55 No.103878672
>>103875808
>the savings
I don't know about them, but where I live, those two cards cost the same used. Very close in raster perf but one is NVIDIA with 8GB RAM and another is AMD with 16GB.
>the savings
I don't know about them, but where I live, those two cards cost the same used. Very close in raster perf but one is NVIDIA with 8GB RAM and another is AMD with 16GB.
Anonymous 01/13/25(Mon)10:32:10 No.103878698
>>103878574
nice
nice
Anonymous 01/13/25(Mon)10:33:39 No.103878717
>>103878669
Catjak is one of the best genners I've ever seen.
Catjak is one of the best genners I've ever seen.
Anonymous 01/13/25(Mon)10:34:46 No.103878730
>>103878647
I never bother with manually detailing or inpainting, which probably explains why everyone else's gens have such better eyes, etc.
I never bother with manually detailing or inpainting, which probably explains why everyone else's gens have such better eyes, etc.
Anonymous 01/13/25(Mon)10:58:23 No.103878942
>>103878730
try it then, eyes are very simple to inpaint and the effect is pretty huge.
try it then, eyes are very simple to inpaint and the effect is pretty huge.
Anonymous 01/13/25(Mon)11:00:05 No.103878956
>>103878730
Manually inpainting is for cucks. Just use adetailer
Manually inpainting is for cucks. Just use adetailer
Anonymous 01/13/25(Mon)11:04:21 No.103879000
>>103878956
changes the style too much retard.
changes the style too much retard.
Anonymous 01/13/25(Mon)11:05:33 No.103879013
>>103879000
You have no idea what you're talking about
You have no idea what you're talking about
Anonymous 01/13/25(Mon)11:07:18 No.103879031
>>103878730
Krita (with an AI plugin) makes it pretty convenient and the effects are huge. It just gives you a lot more flexibility than something like adetailer, especially since you can use it on any random area of the image you'd like. Fixing mistakes by making manual edits (think removing/adding fingers etc.) is also a lot more convenient than opening the image in a separate editing program, then feeding it back into the workflow or whatever. Though admittedly, modern models have gotten pretty good at doing hands/faces, so it's arguably less important than it used to be.
Krita (with an AI plugin) makes it pretty convenient and the effects are huge. It just gives you a lot more flexibility than something like adetailer, especially since you can use it on any random area of the image you'd like. Fixing mistakes by making manual edits (think removing/adding fingers etc.) is also a lot more convenient than opening the image in a separate editing program, then feeding it back into the workflow or whatever. Though admittedly, modern models have gotten pretty good at doing hands/faces, so it's arguably less important than it used to be.
Anonymous 01/13/25(Mon)11:09:41 No.103879051
>>103879031
Krita AI plugin was so buggy for me, all my pony models would overcook and overbake for no reason.
Krita AI plugin was so buggy for me, all my pony models would overcook and overbake for no reason.
Anonymous 01/13/25(Mon)11:11:23 No.103879067
>>103878730
here, gave your gen a new set of eyes and ran one pass on that cross/collar thing as well.
https://files.catbox.moe/rhe78u.jpg
here, gave your gen a new set of eyes and ran one pass on that cross/collar thing as well.
https://files.catbox.moe/rhe78u.jpg
Anonymous 01/13/25(Mon)11:12:56 No.103879083
>>103878956
Too inflexible. What if you want to inpaint something other than the face or the hands?
Too inflexible. What if you want to inpaint something other than the face or the hands?
Anonymous 01/13/25(Mon)11:16:36 No.103879119
>>103879083
Download a different detection model. But that anons problem is faces, which adetailer is perfect for
Download a different detection model. But that anons problem is faces, which adetailer is perfect for
Anonymous 01/13/25(Mon)11:25:52 No.103879208
>>103879067
NTA, but can you tell me what to/how to upscale like this? I'm still not sure how to get defined detail in the image instead of bleeding edges...
NTA, but can you tell me what to/how to upscale like this? I'm still not sure how to get defined detail in the image instead of bleeding edges...
Anonymous 01/13/25(Mon)11:27:26 No.103879227
>>103878730
Use this anon like a pro and ignore this faggot >>103878956
https://github.com/ltdrdata/ComfyUI-extension-tutorials/blob/Main/ComfyUI-Impact-Pack/tutorial/previewbridge.md
it is the only node for you, send latent from first ksampler into vae decode, send image into the preview bridge node, open the mask editor in the preview bridge node, mask the eyes or what ever. Send mask to set latent noise mask, send image into vae encode, send vae encode samples into set latent noise mask, send latent samples from latent noise mask node into second ksampler, prompt for eyes.
Also this
https://github.com/lquesada/ComfyUI-Inpaint-CropAndStitch
this can upscale a masked area then do inference, then downscale and stitch it back in perfectly.
also this
'Open in SAM Detector' from impack pack same as above, can detect just by clicking points in the image, left click blue dots are what you want and right click red dots what you don't want.
I suck at explaining this and I can't find an example of it on youtube, but fucking figure it out, just right click the preview bridge node image and select 'Open in SAM Detector' you will need a model for it.
I might post more screen shots to explain, because there is a a lot you can do, but it would be using way more than pic related.
>>103879013
YOU'RE A FUCKING RETARD
Go back to noob /sdg/ will you and gen slop all day. sick of having to argue with fucking idiots
Use this anon like a pro and ignore this faggot >>103878956
https://github.com/ltdrdata/ComfyUI
it is the only node for you, send latent from first ksampler into vae decode, send image into the preview bridge node, open the mask editor in the preview bridge node, mask the eyes or what ever. Send mask to set latent noise mask, send image into vae encode, send vae encode samples into set latent noise mask, send latent samples from latent noise mask node into second ksampler, prompt for eyes.
Also this
https://github.com/lquesada/ComfyUI
this can upscale a masked area then do inference, then downscale and stitch it back in perfectly.
also this
'Open in SAM Detector' from impack pack same as above, can detect just by clicking points in the image, left click blue dots are what you want and right click red dots what you don't want.
I suck at explaining this and I can't find an example of it on youtube, but fucking figure it out, just right click the preview bridge node image and select 'Open in SAM Detector' you will need a model for it.
I might post more screen shots to explain, because there is a a lot you can do, but it would be using way more than pic related.
>>103879013
YOU'RE A FUCKING RETARD
Go back to noob /sdg/ will you and gen slop all day. sick of having to argue with fucking idiots
Anonymous 01/13/25(Mon)11:29:39 No.103879244
>It's the pony realism retard
You can safely ignore him
You can safely ignore him
Anonymous 01/13/25(Mon)11:31:40 No.103879267
>>103879244
you can use any model, i just knocked something up real quick, I've been more interested in hunyuan video.
you can use any model, i just knocked something up real quick, I've been more interested in hunyuan video.
Anonymous 01/13/25(Mon)11:35:05 No.103879292
>>103879119
But that detection model is again just limited to inpainting specific things. So what do you when you want to inpaint something else that's not covered? Download yet another detection model? Don't know about you anon, but juggling a hundred separate, highly specific detection models hardly seems more convenient than doing a simple polygonal select on some random area of the image and letting the model redraw whatever is inside it.
But that detection model is again just limited to inpainting specific things. So what do you when you want to inpaint something else that's not covered? Download yet another detection model? Don't know about you anon, but juggling a hundred separate, highly specific detection models hardly seems more convenient than doing a simple polygonal select on some random area of the image and letting the model redraw whatever is inside it.
Anonymous 01/13/25(Mon)11:35:33 No.103879299
sam detector in action
Anonymous 01/13/25(Mon)11:38:16 No.103879339
>>103879244
At least pony knows Hasshaku-sama, unlike your holy cow of a noobai.
At least pony knows Hasshaku-sama, unlike your holy cow of a noobai.
Anonymous 01/13/25(Mon)11:39:27 No.103879355
>>103879292
>So what do you when you want to inpaint something else that's not covered?
You can enable multiple, you aren't limited to using one at a time
>juggling a hundred separate, highly specific detection models hardly seems more convenient than doing a simple polygonal select on some random area of the image and letting the model redraw whatever is inside it.
It is, selecting a model that will automatically mask for you IS more convenient than manually doing it yourself, that's the point
>So what do you when you want to inpaint something else that's not covered?
You can enable multiple, you aren't limited to using one at a time
>juggling a hundred separate, highly specific detection models hardly seems more convenient than doing a simple polygonal select on some random area of the image and letting the model redraw whatever is inside it.
It is, selecting a model that will automatically mask for you IS more convenient than manually doing it yourself, that's the point
Anonymous 01/13/25(Mon)11:41:21 No.103879378
should always grow mask to give more context and prevent seems.
but this works better
https://github.com/lquesada/ComfyUI-Inpaint-CropAndStitch/blob/main/inpaint-cropandstitch_example_workflow.png
I just don't want to mess around with it just to prove a point. It can upscale only the masked area making for highly detailed changes, then downscale it back to original size and stitches it back into the original image.
but this works better
https://github.com/lquesada/ComfyUI
I just don't want to mess around with it just to prove a point. It can upscale only the masked area making for highly detailed changes, then downscale it back to original size and stitches it back into the original image.
Anonymous 01/13/25(Mon)11:47:48 No.103879469
i think we should merge back with /sdg/
Anonymous 01/13/25(Mon)11:52:18 No.103879512
le upscaler, flux chuds hate this part ;D
Anonymous 01/13/25(Mon)11:54:49 No.103879536
>>103879512
wrong settings btw, cfg should be 1 - 3, half tile should be activated for seem fixing, i'm in a hurry.
wrong settings btw, cfg should be 1 - 3, half tile should be activated for seem fixing, i'm in a hurry.
Anonymous 01/13/25(Mon)11:55:40 No.103879543
>>103879355
Unless I'm mistaken, each of those detection models is only able to detect a certain category of objects/image areas/whatever you want to call it, whereas manual selection has no such limits. Selecting a specific area in an image takes like what, 5 seconds at most? And you can always be sure that it inpaints only the exact area you choose. Want to only inpaint one of the eyes? Done. Want to inpaint some random object in the background, but leave the rest unchanged? Done. No need to first search for and download another detection model to add to the pile.
Unless I'm mistaken, each of those detection models is only able to detect a certain category of objects/image areas/whatever you want to call it, whereas manual selection has no such limits. Selecting a specific area in an image takes like what, 5 seconds at most? And you can always be sure that it inpaints only the exact area you choose. Want to only inpaint one of the eyes? Done. Want to inpaint some random object in the background, but leave the rest unchanged? Done. No need to first search for and download another detection model to add to the pile.
Anonymous 01/13/25(Mon)11:56:41 No.103879560
>>103879512
ultimate SD upscale works with flux lol
ultimate SD upscale works with flux lol
Anonymous 01/13/25(Mon)11:57:52 No.103879576
>>103879560
yes it does anon, but i don't like flux. If you know what you're doing you don't really need flux.
yes it does anon, but i don't like flux. If you know what you're doing you don't really need flux.
Anonymous 01/13/25(Mon)11:57:52 No.103879577
Anonymous 01/13/25(Mon)11:59:20 No.103879590
One day we will merge back with /sdg/
One day...
One day...
Anonymous 01/13/25(Mon)11:59:45 No.103879594
>>103879577
Oops wrong one
Oops wrong one
Anonymous 01/13/25(Mon)12:00:45 No.103879604
>>103879576
then you'd know a 16ch vae is superior
then you'd know a 16ch vae is superior
Anonymous 01/13/25(Mon)12:02:40 No.103879626
is it possible to avoid "allocation on device" warning, for gguf hunyuan? the first time gens works fine,but not the next time. even with shit resolution lol
Anonymous 01/13/25(Mon)12:23:04 No.103879827
>>103879604
anon he is using ultrasharp for (creepy PEDO) realism, unsalvageable.
anon he is using ultrasharp for (creepy PEDO) realism, unsalvageable.
Anonymous 01/13/25(Mon)12:24:23 No.103879838
>>103877559
All this chart does is blame Indians, Filipino and Taiwanese for the constant degradation of the quality of every single product that Americans make for themselves and for the rest of world.
All this chart does is blame Indians, Filipino and Taiwanese for the constant degradation of the quality of every single product that Americans make for themselves and for the rest of world.
Anonymous 01/13/25(Mon)12:28:37 No.103879868
>>103879838
No, American companies are just afraid of taking risks and they're obsessed with safety. That's why everything they make is soulless. It's time to let go and accept that China is the new king of the world.
No, American companies are just afraid of taking risks and they're obsessed with safety. That's why everything they make is soulless. It's time to let go and accept that China is the new king of the world.
Anonymous 01/13/25(Mon)12:32:03 No.103879901
So is this whatever hunyan thing everyone talks about usable on 8Gbs like Flux? Is it comfy-only?
Anonymous 01/13/25(Mon)12:35:21 No.103879937
I prefer InvokeAI for inpainting over Krita, since it has a lot of AI specific creature comforts (SAM, regional processing, etc) and a good painting UI
Anonymous 01/13/25(Mon)12:40:00 No.103879998
>>103879937
comfy as a backend? looks nice I guess.
comfy as a backend? looks nice I guess.
Anonymous 01/13/25(Mon)12:40:42 No.103880010
>>103879937
This sounds like an ad. Link the best qrd video that doesn't include some retarded faggot's face.
This sounds like an ad. Link the best qrd video that doesn't include some retarded faggot's face.
Anonymous 01/13/25(Mon)12:45:17 No.103880079
Where are the discussions for nvidia's sana?
I want to know whether I can use it for porn.
I want to know whether I can use it for porn.
Anonymous 01/13/25(Mon)12:47:07 No.103880098
>>103879998
Nah they have their own backend I think, it has way worse memory management though lol. Flux chugs on my 4090
>>103880010
Uh this guy has a nice voice I guess
https://www.youtube.com/watch?v=jVi2XgSGrfY
>not sponsored
Nah they have their own backend I think, it has way worse memory management though lol. Flux chugs on my 4090
>>103880010
Uh this guy has a nice voice I guess
https://www.youtube.com/watch?v=jVi
>not sponsored
Anonymous 01/13/25(Mon)12:49:00 No.103880121
>>103879227
ty for the explanation
ty for the explanation
Anonymous 01/13/25(Mon)12:53:01 No.103880173
>>103880098
I want your blue mommy to present the ui to me, as a vtuber in the bottom corner - in that outfit of course. ah I'll stick to comfy for now, forge shits the bed as soon I activate 'soft' inpainting, so that is not an option.
I want your blue mommy to present the ui to me, as a vtuber in the bottom corner - in that outfit of course. ah I'll stick to comfy for now, forge shits the bed as soon I activate 'soft' inpainting, so that is not an option.
Anonymous 01/13/25(Mon)12:55:15 No.103880214
>>103880079
extrapolate from the complete lack of discussion how useful it is
extrapolate from the complete lack of discussion how useful it is
Anonymous 01/13/25(Mon)13:00:53 No.103880290
Furk trying to get the big names to join his discord with special colours
His power has no limits!
His power has no limits!
Anonymous 01/13/25(Mon)13:01:10 No.103880295
I love ran
Anonymous 01/13/25(Mon)13:01:47 No.103880300
>>103880079
You could if you trained it enough. Will it be good? Probably not.
You could if you trained it enough. Will it be good? Probably not.
Anonymous 01/13/25(Mon)13:02:24 No.103880311
>>103879901
Guess it's another fluke then.
>>103880214
There can't be a discusison of a thing if nobody replies to questions about.
>>103880290
Namedropping is cancer.
Guess it's another fluke then.
>>103880214
There can't be a discusison of a thing if nobody replies to questions about.
>>103880290
Namedropping is cancer.
Anonymous 01/13/25(Mon)13:03:02 No.103880320
Anonymous 01/13/25(Mon)13:04:14 No.103880329
>>103880311
I know what a lack of discussion means
I know what a lack of discussion means
Anonymous 01/13/25(Mon)13:05:44 No.103880349
>>103880329
Then you're supporting it by not participating as described. No matter what I ask here, it never gets answered.
Then you're supporting it by not participating as described. No matter what I ask here, it never gets answered.
Anonymous 01/13/25(Mon)13:06:47 No.103880361
>>103880349
correct, so trivially correct that I've already addressed it twice
correct, so trivially correct that I've already addressed it twice
Anonymous 01/13/25(Mon)13:08:03 No.103880373
Anonymous 01/13/25(Mon)14:15:21 No.103881109
>>103873735
>watermark
getting rid of these was a non issue 15 years ago what the fuck are these people smoking?
>watermark
getting rid of these was a non issue 15 years ago what the fuck are these people smoking?