#🍥|anime
1 messages · Page 30 of 1
damn you scared the guy
trying to make this prompt work, even with ClipDrop, is damn hard
a cute girl, white shirt with green tie, red shoes, blue hair, yellow eyes, pink skirt
i was going to make masterpieces
Whats ClipDrop? 
Oh, so like the cutoff extension for Auto1111?
yep, same thing
Nice. Steins?
in comfy
Nice
makima from chainsaw man with my own mix
Ah gotcha.
I may have said "clip drop" abusively though
it's not its name
I'm going to try something™️
New LoCon + Chelodoy LoCon, both are doing funky things that are making the hands better I think
Oo, didn't know about this new feature in comfy
I tried using it but it barely works
it's a custom node : https://github.com/BlenderNeko/ComfyUI_Cutoff
Interesting

Same, I didn't notice any difference
Wonky pic, but good hands
With the two funky loras combined
maybe a skill issue of the extension. using it as described in the repo of it and even tried multiple settings
And again
Civitai Helper doesn't separate Lora and Lycoris 
Chat wants to be stepped on by dommy mommy, absolutely gross. leaves
They don't need to be separated tho
Is that news?
Noooooo, Trexdelussy 
Normal lora can do this?
wth is that
funny letters
step on us while leaving please 
pls dont leave though 
so like the other extension
Yeah that's why I want to separate them
but in text form
you can do it with normal lora my man
Yo deadass finna out you 
Out like a belly button

I'm the best belly button around, you gotta admit
I can?
Oh no doubt
ᵒʰ ʸᵒᵘ ᶜᵃⁿ ᶠᵃᵛᵒʳᶦᵗᵉ ʳᵉᵃᶜᵗᶦᵒⁿˢ
i think these two were caught while hexing

Gnight Smol!
👋
Night Night Smol !
night smol!
keep on vibrating
I just realised how intense those eyes are
I saw that
yeah same 😢
I saw too 
almost enough for the fun police there
dude this is reminding me of bathroom bound hanko kun

hanako*




anyone knows a lora good for holding swords/katana?

🟥 
How much for 30 secs
Probably your life XD

True xD
I try to make waifus. I get blur
mood
OH MY GOD
No killing of waifu allowed
If not you will be reported to the authorities
True art
the dream of any man
no
Where?
So I can avoid it
Just kidding
If it’s such important case, just ping a mod
Or staff
you say they stream pron -> everyone joins to see it
who, not anyone currently streaming is. contact us in #1010934719455707218 with a screen shot and more details please
yeah damn
also @mild tree nice voice
so many people just joined
Tomato sauce
Video proof>screenshots
that was a bait ?
how you know cryonicus btw?
i cant send images there?
what was your previous name in this server?
if it's pron then no, PM me

hehe
nooooo
I need a thicc pillow, if you get my drift
Alt f9 to record with shadowplay to record sus behavior as proof instantly
somehow I do not trust this loli, even if she says its tomato sauce
I was kidding 
seem ok to me, but tired rn
She just has a huuuuge tomato farm
Isekai Nonbiri Nouka
go slep man, self-care most important
what she uses for fertilizer for the nice big jucy tomatos?
Better pillow perspective
well kinda ironic coming from me, considering how little I sleep for recently even though I am in highschool and it's a season with the most amount of important exams
wow
Time for bed fr, nights 

This is fantastic.
Thanks. This is the same model I was posting yesterday with all those more complex illustrations, but now with SAG, the good thing about SAG is that you can transit styles in the same model (if the model supports it) using simple direct txt2img
No SAG and SAG
Thanks
NO SAG?
How does it look that different? 
Are you using -1?
ok, thanks for reporting any porn going around and
all those people joining VC chan all of a sudden
:p
Have a good night everyone
Good nite Guizmoose :O
Research purposes
G'night! 
yeah but I mean, we all have internet here lol x)
Also damn the OP of Tengoku Daimakyou slaps
anime?
DONT LOOK HERE #🍥|anime message

Yes it does change a lot, I increased the threshold (then it keepts the style) and kept the cfg lower
Yeee
damn guizmoooooooose come back I missed the porn 
gn
Yeah but like, for me at 1 it looks crisper, so you are using -1 right?
Yeah I know what SAG is
if you want high quality, i say scale 2 and 1.75 mask is really good
Let me try that 
Nope, I only did this, increased the threshold and left the cfg low (to keep the art closer to the original). In the SAG one I just left the threshold on default and increased cfg to 1-2. If it is giving different it is probably due to the model, or sampler. Sampler is important in this thing
Leg is a lil weird
What threshold 
I was also impressed by the changes, as you can see it is it does change a lot, and this changes a lot of the merging perspectives I had and opens new possibilities that would be discarded in "bad" merges
What CFG threshold tho
CFG 7
oh I know this one
Oh, so just, normal CFG
Shroom lady
i just put on mushroom
idk why all anime models react so strong to those tokens
i dont know her
all anime models do that with shroom girl
mask 1.5-1.75 follows whats in your prompt, it actually generates exact thing, like "roses on water", it will take what else it generated, and it replaces it on the original, its what I have noticed what its doing, it follows whats in your promp no matter the CFG scale, it gives a higher weight
is there good LORA for holding swords/katana?
Dope
probably multiple
i didnt find any in civitai
Thank you ❤️
oh
Yeah, but the original was on UniPC, and I've changed it to DPM++ 2M Karras, then it gets the other smooth style (loses details though), I think sampler is more important than CFG, but I need to test more
SAG has almost no effect on UniPC at least in my tests, I thought initially that it wasn't working
good no?
Fair I suppose :o
thats what comes to mind when i type that
now just shroom
xDDD
even nai reacts strongly to it
@brittle jay so no idea about a lora like that?
like what
holy shit, the difference an upscaler can make :O
katana
oh
damn
i should download one

😉

looks like a cutout i like this 😛
thanks
You forgot I uploaded it @mild tree 
NOOOOO
can't find one actually
You welcome!
I WANTED TO TROLL

I mean a katana holding lora
maybe train one yourself

Good thing I remembered tho, otherwise I'd have to upload it again XD
does it require an op pc?
eh
how much vram you have?
what card?
gpu
why noone in vc :(((( ?
3070
@brittle jayso can i train lora?
How much VRAM you got?
VRAM tho
ram 12
12 VRAM?
Here
hey, should i apply offet noise to my kinda dark loRa?
and if yes, how much?
first one is not used for some reason
Well, I think you can't train
you cant train
I think to train effectively you need at least 8 GB VRAM
i have 8 and thats slow but it works
Yeah
you could maybe train an embedding
12 GB is like the middle spot for training
even with gpu0 having 0 free vram?
I'm gonna make it a new wallpaper engine when I'm on PC again
I mean, that seems to be just your CPU but displayed weird
I assume thats how like K CPUs are displayed, I have KF so it doesn't have display functions
this is a laptop XD
Is it?
ah it seemed like intel graphics card
yes its what its man
Oh, then thats probably why its displayed that way
so no training huh
on that laptop no
time to start an Ai only fans
😂😂
funny thing is, it would actually work
I Dread to think that it actually worked
i think it did lol
93°C 
thats high?
mine around 80 usually
90 when making images
its a laptop its going to be high as there is no good fans
True, I mean
I'm chillin at 50° C atm
And 54° - 60° when generating
Though my GPU does have 3 fans
Well that explains
Mine usually gets around 70C while training stuff, and that is with a custom fan curve that allows it to get hotter so it isn't so loud at night
Also, more dommy mommy queen
its a guy right?
It is not 
i dont know, looks like a trap
Well, it isnt 
Prove it 
I'm using a combination of my Mallyboouwu LoCon & Chelodoy LoCon, first one is for bimbo aesthetic, though its a bit wonky because for some reason it makes hands better, even though its supposedly atrociously trained
Cause I forgot to run the pics through the tagger to give captions
totally a dude
can someone convince me to go to sleep
yep
it's 2am and I have a test tomorrow (today)
poor @magic wave

she is not real
what if I buy a body pillow
it costs too much though

for a long time now I noticed that I love to hug my thick blanket while sleeping like how you would a body pillow
i believe in you @magic wave
Thanks 
so I thought about getting a body pillow
but it's too expensive
unless I can convince my mother to buy it for me
Time to get lewd waifu that you can hug every night
I don't even need a cover tbh
long pillow - body pillow
Its flat so its not the best for sleeping
BUT
Its nice sometiems
Long pillows do be nice
Oh that show exists
I forgot to keep watching it
I watched like 3 episodes and got bored I think
I think I watched 3 episodes, then re-watched Houseki No Kuni again, then forgor about it
except like monogatari and a bit of some other animes
Cause it actually had like, wholesomeness
i havent finished Tokyo Ghoul
It was like a wholesome hentai, with some story
I know it too it's one where they get demon lord out of cat girl
So it was kinda nicer than the usual
that was a hentai?
It wasn't, it was like a hentai, but wholesome, and with some story
Wasn't far from it tho
There was full nipples on the screen if I remember correctly, which means it is not a hentai apparently
Well, I mean, Redo of a healer wasn't either
So

Anime do anime things and go brrrrr
And I do mikoo post

I'm going to try to sleep, keyword try, I might, I might not 
this is me rn
Depends if all my sleepyness completely vanishes when I put my head on the pillow
GN❤️
also pls rate this
hey, how about this style? i creating this right now
i like it!!!

damn how many penis girls do you do per day "(((tranny))), (((trans))), (((trannsexual))), (hermaphrodite),"
That's negative
colors, i like
Also where the prompt I asked 
oh, dont know if this will help you, its a custom model
A cute 1girl, (symmetrical:1.4), face and eyes, __devilkkw/body-1/face_emotions__ face, vector, simple, flat illustration, mascot, cartoon, clean line cartoon, vibrant muted pastel colors, clean line art, colored line art, cute kawaii art, game illustration, pen and ink, vector drawing, emojis, kawaii, anime, centered composition, uni color background, background colors differs from object color
not bad 😛
this happens because of symmetrical, gets it right 50% of the time though
(Tiling:1.4), sheet, in neg
how you think about this? it worth to release?


Maybe I should upscale that one 
I used this one for that image https://civitai.com/models/47067/pikas-new-generation-v10
upscaled to 6144x2048
You making waifus on your waifu now?
if your gonna do it, you go all the way!
https://www.youtube.com/watch?v=UeiZhSadtL0 AMD 😛 Version, and yes its real
Sponsor: Lian Li O11 Dynamic EVO on Amazon - https://geni.us/ovS0F9
The Yeston RX 7900 XTX Sakura Sugar uses AMD's GPU at the core, but straps a wild color theme to the front and goes into surprising depth with creating entire characters around its cards. The Sakura Sugar 7900 XTX isn't just a video-friendly meme card, though. In our in-depth th...
Yo question, are you just generating the images over and over again? Or do you change the prompts/settings at all in between?
adding prompts refines the gen, if you cant describe what you want, redo with a different seed and hope for the best
Yo
Idk if you can help me with this
But I"m trying to figure out how these tiktokers make these videos.
Basically they ai generate, but then the ai generated images, transition to the next set of images.
How do you make a transition from one set of images to the next?
additional data required, link available?
This is the first part of a deep dive series for Deforum for AUTOMATIC1111.
Part 2: https://www.youtube.com/watch?v=nUvPbkAVcZg
SD Deforum Deepdive Playlist: https://www.youtube.com/watch?v=R52hxnpNews&list=PLZmQyPLIePCnmlRpMW-pzXeJGDbjM7moh
In this Stable Diffusion video, we'll be learning about how to install and setup Deforum for AUTO1111 t...
well thats a whole new level of clumsy
Thank you, I'm taking a look at the video now
managed to spill my glass, a can and flip a pizza over and turn the wrong knob on the oven so I burned an oven plate cover
oh and my ultra-wide mouse mat soaked up all the energy drink
except what got on the floor and if it werent for me placing my computer on a chair above the floor
it would've gone right into it
Late nights followed by EARLY MORNINGS! Adventures of a tired and medicated cancer patient part 1: The Spilling.
sounds like a Shudder movie
at least you didnt burn the place down
There's still time
mornin
Im gonna play some This War of Mine, I'm pretty new to it. Any advice that will turn out to be very important to know later?
noble lady showing off her MASSIVE booty
Has anyone tried using different models for inpainting different parts, depending on what it's good at? For example, if one model is good at eyes then one would inpaint with that, and if clothes, then inpainting with that.
on technical aspect, i suspect your initial rendering resolution might be too high
U can make inpainting models using certain models
terribly wrong suspicion
lel
Yes but isn't it okay to use models for inpainting even if they aren't inpainting models?
you sure can, but you'd best use inpainting models when you completely want to replace some aspect
non-inpainting models usually can't replace things, and just generate something out of context
(when denoising is too high)
when denoising is low, you can use non-inpainting just fine
^
I see. Is there a way to search for inpainting models specifically on model sites?
I dont know if many people actively make them
not really, they're... not used often
I played This War of Mine for like 4 hours on 3 failed games
honestely, the better option is: copy image to your favorite image editor, use a noise brush to change what you want
and then img2img that
1 game I got raided by armed intruders with firearms on night 8 and two people got killed on my team
one guy left
If I already have the structure/composition of the image but just want to change its detail and/or colors, a low denoise is ok?
he got depressed and killed himself
yeah, that's what you use lower denoise for
something like 0.2-0.4
chancing colors is hard though, that might need a very high amount of denoise
at least in June/July I'll have some nice offers
so again, better to change the color in image editor
only games I can afford this year will be a few 5 dollar ones or less
for a total of 30 bucks for fun!
Let's say I had an image composition the way I wanted, and then change the colors using segmentation instead. From there, inpainting with low denoise to add detail would be a good idea?
I've still got game pass ultimate though until Feb 2024

Morning everyone 
@cyan burrow massive thanks for the model, its been a blast using it so far, probably not gonna switch to anything else for quite a while either.
I want to make a manga but the coherence and consistency is kind of difficult to manage. However, I think it's a lot easier with 3D. I made a depth map and line art generated from 3d using a plug in. And then put them both in ControlNet.
Depth
Lineart
Output
From here, I need a better way to change colors of things. I think maybe segmentation to change segment colors can be a solution
Is also absurdly versatile for being focused on dark fantasy and gothic
Hats off

I'm so happy with my llm describing the pictures 😄
Maybe dumb question but why isn't everyone using chat gpt? Why local install?
Cool, glad to hear it 🙂
More control, free, fast, and helps develop the open source competitors
Also
also uncensored, you can do whatever you please 
I kind of left the dark fantasy realm for 3.0 and focused on more versatility
no censoring
You can prime local models a whole lot better
I wish more people would pay attentionto my model though, it falls into the shadows on civitai
chatgpt is honestly very good, but its: cencored, and is always in a default state
mine already has 😄
and while I love the shadows, I don't want my model there
I see. I was thinking of giving a chapter summary of a story and then it could write an entire chapter. Then I could do that for each story chapter. That way it could write my whole book for me. But I haven't tried any chat bot at all yet
some people release a model and within 5 minutes they have like 10+ reviews and 5000 downloads
@terse ore you'd be surprised how competent it can be when given the right instructions
Very fucking sad yeah, gonna try lora compatibility today, plus some more pastel style stuff i think, then i have enough to write a comprehensive review
Already proved enough that it has PERFECT synergy with the bf3 lighting LoRa
maybe needs more booba to lure people in
give it the instructions of "write a story about a man stuck in two realities, shifting his consciousness back and forth and caught in a war on both sides, what he doesn't know is that he's fighting his own faction with his own faction on both sides from another temporal perspective. He is his own antagonist"
NOOO
fuck, i'd first have to write the bot 😦
I'll have to try it sometime. I have a decent amount of RAM so that could help getting a beefy one. Not that it means it would be better. Vicuna is light but apparently better than many
No boba pls
I am so happy this model isnt horny
vram. and for 13b models you really need 12gb+
Doesn't llama use RAM?
Its just so sharp
and... you can't use the really big models either
Clyde blocked my other post 😢
Judging by your name, that's probably a good thing
I try to keep em tasteful for this server
I'll just use chat gpt for now and then see what can do with the other ones later
I wonder, are there any good examples of manga made with AI in its workflow? I've seen a few comics and they're...not great
probably hard to get consistent character
That is, they don't have much physical character interaction
I thought Loras/TIs are consistent?
@terse ore total abuse of llm+sd
the medic
William, also known as Fenberenr in another reality. Grew up shifting between two lives, going to sleep in one was waking up in another. Every moment except sleep accounted for. As William, he fights daily to protect his family from an invading force of demons that have infected the population with their conscioiusness in order to pass the rift of realities. Resulting in the zombiefication of the host and the unintended corruption of their soul. The response has been for William to create a technology of machines that pass the quantum entanglement and force their way through the plank constant to close the rift. The result is the shift of random locations into a realm of darkness with creatures of nightmares. Also unintended as the first intention was to close the rift.
Both sides now ending up in a full on war, William and Fenberenr are the leaders of the ultimate fighters and forces on both worlds and the ones to inevitably come into conflict. With the temporal differences in the nature of the two, their inevitable conflict as they clash will crack the universe like an empty egg, leaving tiny pieces floating about the chaos realm

That's like using a color pallet? I think ControlNet has something like that
that would be me being a storybot
Holy shit
my brain that is
no, the LLM just made it all up
the story of the protagonist being his own antagonist
Can it generate panel to panel consistently?
No medic 😭
whoops, left picture mode on xD
but my first story is different, about a man waking up after blacking out and finding himself reincarnated in another world but missing a large part of his life. His consciousness also split into two, though his mind acts as one. His new body is a psychopath, so he only has his old moral code to go by. Which means very little after he gets pulled into a Lovecraftian story about the guidance factor of the universe, the lines of fate and barriers of freedom resulting in the protagonist declaring war on the Universe
its a bit harder to explain that story since it will be a storyarc based one
with the intention to last awhile, milk the cow
I just threw a small bot prompt together, so output isn't high quality yet XD
that is way too Kang the Conquerer
I think it's possible to make a consistent manga with AI. I've paneled everything and written the dialogue and everything. I think I'll use rigged 3D characters instead of Openpose. The SD model will be a cel shaded type instead of a more complex gradient, as that will make it easier to edit in an image editor
it is possible but you need to make embeds or loras
cel-shaded just makes me think, why not just use 3D rendered cel-shading?
There is also layer separator which will segment the image and make each one a layer in Clip Studio
I also thought that but 3d characters retain their composition. So I just need to color correct them
3d cel shading looks pretty bad. You can tell it's 3d
His outfit is complex. Does it change from generation to generation, or can you control it?
it'd be nice to get to Guilty Gear level cel-shading
Vicuna 13b
speaking of manga, I don't think I've seen many black & white loras /checkpoints
like nobody actually trained on manga panels
all on highres anime
This is the best I could get cel shaded in 3d, but it looks wrong still
but then I can always recolor it in photoshop
It's the same character I posted above
Do you need it less or more 3d?
You can always put the opposite in your negative prompt
(that's what I do at least)
Yeah that's my plan too, recolor in an image editor. How about backgrounds? Can you make those consistent?
Less is better. I don't need 3d shading but I just use 3d for the structural consistency
but I can always run it through clip studio paint pro or photoshop character recognition
where it will take the character out without ruining it
Yeah the eyes and hands usually need work.
then I can put it on any background
The current llms have a context memory of 2048 tokens, so to get it to do stuff consistently remains hard.
Ah that makes sense
It can remember a bit, but after a while it forgets the oldest memories
Am I high or smth
Nah, just me casually breaking reality
I use 3d to make consistent backgrounds and then need recoloring, but If you don't use 3d, I think one technique is to use Google maps and then use ControlNet and a preprocessor. That way you can do basically the same thing
or unreal engine and build it yourself
its royalty free up to like 100 000 or 1 million
I still need to try multiple characters in a single image. Maybe multiple 3d characters in a single scene and generate, then use segment and change colors from there.
I have a few hundred GB of 3d assets from...sailing...
strange waifu coming from you 😛
you can always convert those into unreal engine stuff
Building scenes takes time though. But compared to webtoons/manga authors doing a ton of manual work it's fast
she's in there, piloting
Is there a reason to use UE over blender or other?
it's free, the story contains a ton of free stuff, it requires no royalties, it has near pure photorealism
so for backgrounds you can make anything
anything
with preset fabs too
then you run them through img2img or control net to get anime version
I see. And one would feed it into ControlNet I assume?
Yeah
sure or you can extract a much clearer depth map frmo UE5
and feed that into control net
since that depth map would read the light at a LYDAR state
LIDAR*
Yeah my depth maps are pretty trash. The reason I use Maya is because it has Pencil+4 which is amazing at extracting line art
I remember when Maya was new on the market and just a start-up competitor vs 3d max
look at it now
Psoft recently came out with Pencil+4 for Blender. It was used in some new evangellion I think, before it was out on the market. The guy said it allowed him to keep it in a blender workflow which saved time
I think Maya is losing our to blender because it doesn't really adapt or change much
I'm waiting on AR/VR combo that features AR sculpting
so you can sculpt at real size
It's been used a lot because people can't switch into other pipelines. I think blender adapting for AI and Maya not, will have people moving to blender more and more
the tiniest detail
Maya will fall behing if they ignore AI
Adobe and I think Cinema4d are embracing it too
Yeah they are ignoring it as fast as I see. It's just hobbyists making plugins
clip studio paint is throwing AI in the trash (they should throw their healing brush in the trashthen too)
and Krita really needs a AI fork
Adobe are making their own AI tools
Firefly etc
Yeah I was so disappointed when CSP scrapped AI
@wispy canopy there she is
I was going to lay for their 2.0 version but because of that, I'm going to wait until it comes out for free
hopefully CSP will see the error of their ways and recognize that they don't need to embrace disrespecting an author saying "no don't use my art" to make use of ai
inpainting? 😄
or new image and copy paste?
yeah lol, I need to make a mecha cockpit lora I think
I bought their version 2 for 20 euro
with version
1
I bought V1 back in December, with free V2
it's impossible to describe well
actually a new image would've been a smart way to go
and just paste it in
but in any case, hard to get a mecha pilot
agreed
having a generic model of a mecha with a colored vivid pilot cockpit and seat
with it closed, open, broken etc
that way the AI will recognize its placement without forcing a mecha, then make mechas in another Lora
I think Eva will have a lot of good cockpits
then merge the Loras
Japans Government has sent her a bunch of health letters. She ate em
Maybe now with ControlNet getting better and better, reference images can be better used to make specific things in different styles
lol
but I would love to make most of the stuff in blender in 3d and then run it through AI
what is specific tag used?
cause I think there are speed sculpt methods
chunker
and then auto skeleton
I haven't used latent much. It changes and mixes things apparently. But there is an extension, I think it was regional prompter. I need to experiment with that stuff after I'm done testing a few more 3d things
latent couple with twoshot etc you can just paint the areas you want to prompt
There is a blender rig that has openpose, hands, landmark face, mediapipeface, canny hands and feet, depth hands and feet.
for what you want in there
for blender I just want to learn how to sculpt fast, different methods
Hmm. I wonder how well it can do in physical interaction. Like if a character grabs another by the collar, would it make the cloth folds and hands around the cloth and so on?
Because when you paint, it makes a blur
My friend
Why
Why are you cruel
What did they do to deserve that


uncanny valley vibes here XD
Fair XD
shame about her hand
the eternal struggle
gimme 5 minutes I'll fix it
here, to cleanse your eyes ❤️
any tips on how to replicate this pose?
Oh thank you, i was about to go blind
Controlnet (maybe)
whats a controlnet?

my fan is really spinning now
the horror
plz don't report me for those abominations ❤️
❤️
latent couple adds a lot of time to these gens
they be copying my red hair!
Kinda XD
DANGEROUS C.U.R.V.E.S. AHEAD
the thickest thick anime girl in the history of thick anime girls
nah, we can go thicker, it seems
make her american
,....photoshop crashed
first imma install cut-off tho, it seems to be mixing elements
fuck I hate crashes
this is a thicc unit too
Thanks
thanks
WE HAVE GONE TOO FAR. BACK BACK BACK
you have turned this chat to /d/
I just post bimboe 
I think I will download hands from clip studio store
I was trying to do princess rosalina, seems the model doesn't know it, but now I have a new plan 
GN
any recourses on how to install SAG?
Its right there
GO FAT! the NEW HOT!
https://civitai.com/models/48197/obese-girls-or-concept
A request For generating really heavy girls Works best with deliberate tagging, also trained on a few keywords but they dont seem particularly pote...
You made dat? 
and i downloading the files into the scripts folder?
Wwwwwwwwwwhat
im sorry...
Hit this
mhm
OH
Install
ofc
here's a hand to use, rotate and scale into place
and it can be used for gripping clothes or holding something
ye
And run it, and thats all
Then you restart your SD, so close it and start it up again, and its good to go
ok thanks

what this even mean ?
highres ❤️
what did i do wrong
you could paste these over images youve got issues with hands
remove the hands and color in a bit
and then img2img
different hands for different generic grips
want more?
Install from URL
https://github.com/ashen-sensored/sd_webui_SAG
Install
Apply and restart UI
close webui
.\venv\Scripts\activate
git checkout 22bcc7be
git apply --ignore-whitespace extensions/sd_webui_SAG/automatic1111-CFGDenoiser-and-script_callbacks-mod-for-SAG.patch
launch webui
both of these lines? in the cmd root directory
yes 1 at a time
🤔

ok
did you activate venv?
no...?
type
git checkout master
git pull
git checkout 22bcc7be
git apply --ignore-whitespace extensions/sd_webui_SAG/automatic1111-CFGDenoiser-and-script_callbacks-mod-for-SAG.patch
from this env should be venv
EEEEEYYY
it works
thanks
sorry for bothering you guys, i just dont know anthing about the technical stuff pretty much at all

whooo, fancontrol for the win ❤️ gpu runs cooler, and gpu fan spins less, offloading cooling solution to the casefans -> More cold air being pumped in = less need to run the gpu fan at full throttle
Is it free to use ai tools of these services ?
O.o
Most of us run it for free on our own PC using this https://github.com/AUTOMATIC1111/stable-diffusion-webui
:O
you going all in on these huh? chubbychasing?
if your gonna go in, go all the way in
@cyan burrow
CONTEXT PEOPLE!
lol
hmmm... my randomizer isn't random enough anymore
injects more random into his veins
and my new gpu + cooling setup works a bit too well... i've heated my room 1.5 degrees without having the heater on...
this summer will be hell 
no need to speed it up XD lol.
hey, try absolute unit, my model gives me huge lols when i prompt that
well lets see
absolute unit, chibi, dinosaur suit
nice
it's pretty obvious there's some training done on that specific term
"In awe at the size of this lad. Absolute unit"
looks like the mc from Tenkuu Shinpan / high-rise invasion
#1 tho










