#๐โai-images
1 messages ยท Page 1 of 1 (latest)
hey
ok
top 10 inglesis capturados por camaras
Shouldn't be just Ai-Art Since Dalle-3 is a Thing?
poo
I shall restart the diffusing
@tawny peak has reached level 1.
Ilaria was here :p
bruh
I think just give the mods and admins some time to rebuild this server, and they might change it
and Dalle is just a specialized version of Stable Diffusion
@uncut robin has reached level 3.
Yo!
Since the old server got nuked, lemme repost some of my stuff :3
Nice, that actually looks like a proper anime character
damn
anyone has novelai old model ?? want to revisit since I learn AI from that leak (having recent one even better), feel free to DM if it's prohibited
Can you still do stable diffusion with colab?
@wet rune has reached level 1.
I got it from a torrent link Iโm not sure if it works anymore
that's where I got the model too, unfortunately the torrent seems dead and finding it publicly is impossible since they pretty harsh and fast with copyright
some of my old renders
which client you use?
automatic1111
my newest generated image
I wouldnโt drink that tea
She just says that itโs a body fluid
It could literally be anything
It could be spinal fluid
This should be renamed to ai art
Agreed:
On a side note: Using Free Midjourney is fun but is a pain in the ass to wait for
Could be better...
Im switching between Stable Diff XL and Midjourney lol
Its so diverse in terms of styling its weird
Though in stable's case, album covers are king

is there a colab for stable diffusion?
Doubt
nope
i mean yes but u have to pay
Though one of the earlier Stable versions was more accurate... but I think that was just a lucky guess lmao
This was under 0.9 btw ๐
Along with this.
and I believe this too.
@uncut robin has reached level 1.
but again I can't tell which ones are 1.0 and which ones are 0.9
I wonder if Midjourney can make wacky images too...
Ooooh
I mean Stable makes cool images and all... but I wonder if Midjourney can do the same?
wat?
I wanna try one of your prompts on Dall-E 3 just to see how it turns out
it wouldnt work
cause of copyright lol
one part of my prompt process is litterally placing the names of the Celebs.
Man just like a day ago it wasnt this strict
Which is sad to see cause I wanted to see them in Dalle-3
At least I can try again in Midjourney....
first image I literally made was.....Gordon ramsay eating spaghetti with his hands..
but realistically, Midjourney (Free) takes forever LOL
I dont get it though because theres like things in the image that can prove the image is ai
Like rn its still processing my previous prompt
Midjourney is free again?
Like has a free version
ClipDrop still ontop
Honestly I rather stick to ClipDrop for now...
but then again Im not a Midjourney Prompter.
I used to love MJ back when there was that infinite free trial thing
Oh yea fr
The one thing about ClipDrop is the fact that prompting is so easy
you don't really have to think as hard as midjourney.
Wtf
Who are these people and why are you so obsessed with them
I mean... One moment you can have them drive away in the Getaway Car, and another you can have them hangout with people like Sleepy Joe over there ~ Donald Trump
Do you ever squint sometimes thinking this was Katy Perry or Carly Rae Jepsen?
Who TF are theyy
What do you mean youve sent like 100 imagen of them
Like look at these blokes
...
Oh yea fr
Theres literally a tag in the images that say theyre ai
Then come all the pay to use sites
tru
Another getaway car moment
lol did I mention to you guys how big my album "covers" are?
Like fr, I have a HUGE folder of all these babies
Clipdrop aint even that bad, its just bad in like....doing more than two people
@carmine vault has reached level 2.
Im dying to see what XL 0.9 is like though
and correct prompting
Well... they used to have 0.9 as a standard for ClipDrop
then they bumped it to 1.0
I wonder why though...
wouldnt 1.0 be better? I dont know why 0.9 is a paid thing
Well that Idk
but I know when ClipDrop was first introduced, it had an earlier version number
0.8 I think?
This is the negative prompt I use for everything in ClipDrop
ooo
Well for me its all this:
(deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime), text, cropped, out of frame, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck, BadDream, UnrealisticDream, 6 fingers, 4 fingers, distorted fingers
shiny skin, smooth skin, out of frame, (bad fingers:1.2), (bad hands:1.2), (bad limbs:1.1), (bad proportions:1.2), (bad quality:1.1), (poorly drawn face:1.3), ugly face, (poorly drawn:1.2), asymmetrical body, (bad anatomy:1.3), (flesh merge:1.2), (large breasts:1.0), (fleshpile:1.2), tiling, (generic:1.1), glitchy, (gross proportions:1.2), (malformed:1.2), extra arms, extra fingers, extra hands, extra legs, extra limbs, extra body parts, extra breasts, mangled fingers, (masculine:1.3), (bad body structure:1.2), badly drawn hands, (badly drawn:1.2), (amateur art:1.2), missing arms, missing breasts, missing extremities, missing hands, missing limbs, missing lips, mutated hands, (mutation:1.2), poorly drawn hands, asian, abstract, (anime:1.2), (blurry:1.1), cloned face, (deformed:1.2), (disfigured:1.2), double torso, fat, long body, penis, skinny, testicles, , ugly, cropped, jpeg artifacts, signature, blurry.
Essentially I copied all of it from a stable prompter that is supposed to be realistic
for Humans at least.
Sometimes it works sometimes it doesnt
Eh, the thing about that it has negative things like "cgi" "sketch" "drawing"
my negative prompts are all basically like blanket prompts that I can use for anything
You're playing with fate here xD
also: another Getaway car moment
Now that is more dramatic in 1.0
Kinda seems real from a distance.
Bings AI can really detect the names too, like I tried "GordoM RaNsayY" just to confuse it a bit but still blocked lmao, they must be using AI to see what the names sound like
theres nothing even in the code of conduct about it
Thats true
but I think its for copyright reasoning
not using likenesses or whatnot.
But here's ClipDrop and Midjourney not caring about that shit
@uncut robin has reached level 3.
LOL
omg
the only thing I see is this which like fair but
i dont like gordon ramsay eating spaghetti will cause mass destruction, theres already anti-gore and anti-nsfw things in place
I mean if it was true then other AI generators would follow suit
I guess you could say Images have a few valued wordings
oh gifs dont work here
Yea they disabled it
Looks more like BJ Novak to me LMAO
How come?
Cuz youtube was trying to resize my first video lmao
Which is why I had to stick to 1:1 ratios
Man I gotta start using ClipDrop again, the one thing I hate is the watermark
like atleast with DALL-E 3, the watermark got removed when you downloaded the pic
why else did you think I have like 300+ AI images
of these 2 blokes who people dont really care about
Ok (Well her Swiftie Fanbase yes but still)
What do you use to remove the watermark again? I remember the website but doesnt that have a limit?
ah yes
I use this one https://www.aiseesoft.com/watermark-remover-online/#
Its unlimited and free
Ok how am I progressing down? LOL
Thanks man, bookmarked it lol
Im doing a little comparison between the two right now
Well thats cause there wasnt a whole lot of people who know of ClipDrop
and now there's a fuck ton of people XD
4.0k at its peak now .-.
for queues LOL
Oh damn
This is Dall-E 3
Clipdrop
Honestly cant tell which is better lmao
I mean it was the only good one out of 4
this was another ClipDrop one
the prompt was just "A thumbs up on a plain white background"
I mean all the D-3 ones look solid
how do i do this
lemme try on automatic1111
i aint staff uwu
use bing image creation
it uses dall-e 3
ok
google it and ull find it
ok
you get it, no worries :D
thank
DallE has a pretty cute artstyle sometimes
@neat nexus has reached level 1.
Is it DALLE3??
Yes
Wtf is thos chat....
@halcyon pivot has reached level 1.
its generated ai art?
AI Art that is generated
Like that sentance
It was generated by KidChatGPT
Also here's a wedding ring for when Taylor Swift marries that one Football star

and here's a lost album cover of a cover
Can I say Taylor Doo and Billie too? 
Now this is quality art.
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
.
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
@uncut robin has reached level 4.
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
Imagine trying to Folder-Dump your entire catalogue of AI generated art
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
Vote Booster: 10% (11 hours remaining)
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
U almost there
XD
jesus
and im not even done
Could spam images I here
This entire catalogue of images of my AI creation
its so huge people think im crazy XD
Like theroically
I can get to level 5
by showing you guys ALLLLL
of my images
Whew j would do this too
I have enough ig
oh and if you're wondering, Its Billie Joe Armstrong and Taylor Swift.
XD
or at least its supposed to be*
Well Taylor swift I do know
Well yea, her "style" is so pop culture now
that no one can forget her.
Ever
She essentially is immortalized into the pop-culture zeiguist of today.
Everyone spamming in general with slow mode and this here feels illegal
meh
not really, Im just showing off my AI generated art lol
from my massive folder xD
Welp I need my SD creator role back
I know the feeling lol I have about 500 images and that's after I edit most of them and deleting trash
oh and thank you for reminding me about the extra images I used snipping tool with XD
Fuck watermarks xD
Lol
Remove tool to the save ๐
I already do have the tool for it ๐
Its called
Unlimited: No bullshit, just good old removing the watermark tool
Thats how I was able to make album covers XD
@uncut robin has reached level 5.
Congrats
Level 5 and I Still have more to share
All these are my early album covers btw
Vote Booster: 10% (11 hours remaining)
โจ Tip: Use /card to modify your rank card
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
CONNOR STOP SPAMING
oi, arent you supposed to be playing games with trump and obama?
XD
We broke Up ๐ฆ
Theyy say i'm a noob but idk whats wrong with countering pharah with reinheardt tbh
thanksssiieeesss
crisis averted
Can I get my SD creator role back too?
u didnt have it
but i like ur stuff so sure
Welp I did have it
Got it from Sam
Idk why this isnโt ai art channel like the old server
And thx
i dont rember though
Got it a week ago
gave u the roles back welcome
Ty
Hello
@waxen arch My pfp was made with AI by the way :)
Ok
mine is bigger
amateurs
A little less than you
(This is mostly Bing Image creator images)
well i have an rtx 3060 ti, and token merging gets me below a second/image
@peak siren has reached level 1.
Nah
ah wait until i get the hotshot xl gif repo working
it simply refuses to use my gpu, but it loads everything into it
they are working on an extention for automatic1111
What GPU do you have?
@tawny peak has reached level 1.
rtx 3060 ti
oh yeah forgot levels reset
huh, I could try it to see if it works I have a 3060 as well
https://github.com/hotshotco/Hotshot-XL
well, but you need to download a 17gb diffusers model, not quantized
That's fine I got the space
i recommend
python inference.py --pretrained_path="modelpath(diffusers)" --prompt="a flying cow, hd, high quality" --output="output.gif" --precision="bf16" --xformers
it was running on cpu 13%, 35s/it
i have no idea why its not using my gpu, i even changed something in the code for using cuda:0
Idk, early software issues maybe
in ai terms its slow rn
I mean the gpu issue not the slowness
i asked the developers for what was wrong with it, but it seems they are too busy
@peak siren has reached level 2.
I mean it did just come out 2 days ago
yes that's the crazy part
also had to update diffusers package to make it run
but the usual pip install -r requirements.txt worked just fine
I'm not too technical with python so where it says: virtualenv -p $(which python3) venv, what do i put for the python3 like the version or the path?
ah no ignore that just install it normally, or as i did with, and the command works with
alternatively
python -m venv hotshot
hotshot\Scripts\activate
There is no need for an enviroment, if you are not developing it with requiring of an exact package version
I thought they wouldn't have a complete repository so I thought I'd just try the setup instructions they had. But Cloning works
@tawny peak has reached level 2.
yes just make sure you get the path
and if there is an error, do pip install diffusers -U
yep it's all good
Just have to wait for my slow internet to install it all
Although most packages it requires i have cached
Well I hit one error during install already, it tries to install a version of cudnn that doesn't work with the installer, so I'm just going to try a newer version. It probably won't work considering I've tried a similar thing before and It didn't work
they released a safetensors model for the temporal layers, no idea how to use it
Me neither
oh, that's just compiling cuda
yeah and it throws this error using requirements.txt
as long as you had it installed normally, there shouldn't be a problem with that version
I guess i'll try
Dick measuring contest except instead of inches/centimeters it's in gigabytes
I think Hotshot XL only works on linux as of now, as it requires packages that are linux only
Well works fully on linux
I unfortunately can't get it to install right now, I could try with WSL but I may just try it when it's actually fully functional and at least a few weeks developed
a few weeks ๐
wow are they already 15 XD
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
โจ Tip: Use /card to modify your rank card
hi
@waxen arch has reached level 1.
Vote Booster: Vote now for a 10% boost. https://arcane.bot/vote
how to use stable diffusion
@agile wraith has reached level 6.
Some know how do a dance?
Raiden Ei(Genshin) finger alwasy sucks
Twins?
Also left girls ankles are definitely broken for the feet to turn that way xD
Ohh ai hub has a Mascotte
Wait you made it with only the 2 pictures?
if that's so, i didnt think loras would be that powerfull to train, i trained one with 30+pictures and it rendered great
What images were used for the lora?
YES LET THERE BE ART
agreed
time to generate some more art
i wonder
whos fursona shud i make a model of next?

can u maybe do like a pfp of my little husky doggo but holloween themed?
wanna do some dumb stuff
โจ
is your pfp the only art you have? :o
<w<
oi
its all AI generated
i dont do art lmao
i suck at it
a bummer ik
<t:1529880598:F> - (<t:1529880598:R>)
<t:1696493049:F> - (<t:1696493049:R>)
litsa_the_dancer
460577350900514837
#6677cf
<@&1159293140440723499>,<@&1159293204038955078>,<@&1159332519167144039>,<@&1159337831928434799>,<@&1159338005404844083>,<@&1159344375290138665>,<@&1159345504442925056>,<@&1159434617451978802>,<@&1159520205395984475>,<@&1159579982683701379>,<@&1159670353333854208>,@everyone
u can download the image from here and use clip ig
o-o
i mostly work with loras XD
but heck i guess i can see what i can do
even ai generated art is fine ya know XD
Lmao good luck
And thx
first of all
@quaint tiger has reached level 12.
this shit look cool af
Oh I like it give me the prompt
I will also do some tests myself
i used img to img
but heres prompt
Husky, Male, Furry, fullbody illustration, adult, purple hoodie Hoodie, black and white fur, white spot on forehead
(solo:1.3), head portrait, by haps, Cory Loftis, Yuumei, by braeburned, by Tom Whalen, by Shiitakemeshi, adorable, (male, ๐ฅบ, blep:1.2), (detailed expressive face, pleading, playful:1.2), detailed fur, solo focus,,( detailed colorful fur), hi res, masterpiece, cute, lovely, clean abstract background, shorts, sfw,
negative:
boring_e621_v4, (watermark, text, signature), (blurry, blur), (worst quality, low quality, lowres, low res), (deformed, distorted, disfigured, distorted face, distorted eyes), bad anatomy, ugly, extra limb, missing limb, floating limbs, disconnected limbs, amputation, extra tail, missing leg, monochrome, sepia, (Dutch Angle:1.5) (NSFW) (Orange fur)
Do u just come up with these wtf
I need to improve my skills lmao
Does it have zeta haru style maybe
zeta haru?
Dunno him?
nuh
Ohh come search him up rq
Ok I'll try some Halloween stuff then
Still using the yiff thing Right?
yes XD
Okay well at least it has a sfw option
So it will work just fine
how i can do a dance whit sable diffusion?
Oh that's hard
But if u go on Google and look around u might find something
Mhm looks good but I think I'll opt for this artists style
nah y try but is not working, so meny errors whit no sanse
Go onto stable diffusion og discord
They'll definitely help u
theres so many styles you can go for
What's the artists name
which one?
I know nothing about furry artists lol
This one
by braeburned, by Tom Whalen, by Shiitakemeshi
Ohh it's a mixture of em.
mmmhhhhmmmm
Kk I'll keep ya posted . I think 512x512 is good no?
And should I use high res
Yeah I see
Wow that low
It's to avoid doubling n stuff right
those are the hires steps
But at the cost of details
Ahh I see
Ok will do some basic chemistry in my AI laboratory and I'll see if I am able to come up with something spooky ๐ฟ
with your prompt XD

lmfao
it looks so weird
it just like random
i got the idea from this
likke useing adversial prompts and shit
interesting
it does work for LLMs so i wondered what about Image Generators
Yeah I don't really seem to like the zeta haru style here
oh i didnt use their style
this one was the closest i cud get XD
walter
he looks like that rich old guy from movies
how does one stable diffuse?
i know how to dalle
Shhhhhhhhhhhh
i using my 980ti graka and run it
how get that
ok
Oh hello
Furreby
my dumbass thought that triangle was a play button
@hybrid ingot has reached level 5.
lol what triangle
cant see
in the first one
nice
i will upload here sometimes cuz i have too many :welp:
lmao
hey hubby.
NotImplementedError: No operator found for memory_efficient_attention_forward with inputs: query : shape=(2, 4096, 8, 40) (torch.float16) key : shape=(2, 4096, 8, 40) (torch.float16) value : shape=(2, 4096, 8, 40) (torch.float16) attn_bias : <class 'NoneType'> p : 0.0 flshattF is not supported because: xFormers wasn't build with CUDA support Operator wasn't built - see python -m xformers.info for more info tritonflashattF is not supported because: xFormers wasn't build with CUDA support requires A100 GPU Only work on pre-MLIR triton for now cutlassF is not supported because: xFormers wasn't build with CUDA support Operator wasn't built - see python -m xformers.info for more info smallkF is not supported because: xFormers wasn't build with CUDA support dtype=torch.float16 (supported: {torch.float32}) max(query.shape[-1] != value.shape[-1]) > 32 Operator wasn't built - see python -m xformers.info for more info unsupported embed per head: 40
Time taken: 1.5 sec.
do i need to reinstall xformers?
my husband is busy atm
A100 GPU ๐
I think a100 is much better tho
I'm not 100% sure, but u can try either reinstalling xformers or add 'no-half-vae' to ur bat file
ohh yeah i can try that. im acyually thinking of use a baked in VAE instead
should help me avoid errors n such
Welp, not sure about that too, but I think it's not related to this specifically but more to half precision and full precision of the vae
mhm i see is there any way to change that
probably baked in ones already have that sort of stuff pre detrmained
I'm not sure myself so I can't really say, but it's just what I think it is since using SD for quite some time and seen too much error ๐
lmao i get it. im fairly new to this lmao
but im sure i'll fiddle around
thx for telling me though!
Im not on my pc so if u can't fix it just ping me here if you want and I would try to find the reason
mhm im trying with the baked vae atm
why am i not allowed to post any images anymore?
we had to turn some stuff off bc of viscous attempts of our server being taken down
well that sucks and makes the server feels dead to be honest :(
yeah but we dont want to be taken down again so bear with us for the mean time
also can you tell me how i could get a SD creator?
post frequently and i give to ya
when you would know how much my playlist on youtube has xD
what the fuck XD
also why you callin me ya hunny a silly butt
hubby*
why not
XD
also i have never seen that error before
thats good atleast :P
yes i got it resolved by the experts in those stable diffusion servers. its an xformares thingy
that messes with the python ver u have
"What we do here is go back back back" 
yep I also switched to torch 2.1.0 with cu121, at least 1 it/s more than with xformers
always great to hear 
@shadow bane has reached level 1.
yes baby dw i didnt forget about ur promotion
aye thank you! Slowly getting back into things here 
Oh my, just went through this chats history... Connor really loves Taylor huh? 
yup
are image uploads now gone for everyone, or just new accs?
aayyyyy ur back!
gone for now
yeeeeaaah! I was super busy last few days, so no time to rejoin xD
sad, but hopefully they can figure out a system that keeps out malicious actors but allows normal users to share 
I want to post so bad lmao
I would but no image perms
actually no I wouldn't either way
Can any mod give me image perms in here
no one have image perms here
yea
they only can give perms for embedding
sad
no can do we have some problems atm with some ass holes trying to snipe the server with copyright
so yeah we cant
so sd creator role give u perms right?
nope
this wont give me perms

@quaint tiger hey babe I lost the link to controlnet mind sharing again thx
Too lazy to go online that's why so yeah
Thx๐ถ
but they could make it only for this channel to be honest since SD creator is not a automated role ^^'
I would think that's a good idea 

SD GOD
We might as well
yes please ๐
what did u use for this to make it very similar?
add (by Zeta Haru) to the prompt
no look at the one i pinged
u sure?
it brought me to the wrong ping the whole time
yup it didnt load
XDDD
also i used img2img controlenet for that
yeah which algo?
i drew ๐
dunno how it works though smh
dropping this cuz i liked it
mhm
should i choose all the control types or nah
i only have a canny model installed
1 moment
sure
i did something
i gave it this
Husky, Male, Furry, fullbody illustration, adult, purple hoodie Hoodie, black and white fur, white spot on forehead
(solo:1.3), head portrait, by haps, Cory Loftis, Yuumei, by braeburned, by Tom Whalen, by Shiitakemeshi, adorable, (male, ๐ฅบ, blep:1.2), (detailed expressive face, pleading, playful:1.2), detailed fur, solo focus,,( detailed colorful fur), hi res, masterpiece, cute, lovely, clean abstract background, shorts, sfw,
and it came out with this
negative:boring_e621_v4, (watermark, text, signature), (blurry, blur), (worst quality, low quality, lowres, low res), (deformed, distorted, disfigured, distorted face, distorted eyes), bad anatomy, ugly, extra limb, missing limb, floating limbs, disconnected limbs, amputation, extra tail, missing leg, monochrome, sepia, (Dutch Angle:1.5) (NSFW) (Orange fur)
yes i have them saved
oh
so goofy
this is cute though
i like it but why isnt like copying the hell out of it?
of the image given i mean
simple
denoising strength
im using a 700 mb model btw
on hires?
my gay ass is in txt to imgae
i thought you were using img to img XD
wait let me crank some 90s rq
but what do you want different then?
the style
sd does it much more clean
by haps, Cory Loftis, Yuumei, by braeburned, by Tom Whalen, by Shiitakemeshi
this
replace it with other artists
if i do breaburned only
?
is that ok?
so try by dagasi for example
they need to be artists from e621
cuz that model uses e621 database
want a list of artists?
sure
yeah brea is on e621
ok ima dm it though
hehe
ok i get it
blep
yes

yeah I get it, it's just so funny/cute seeing it written out 
x'3
derpy and massively cute 
yeah bro looking cute

nice husky bro, I wanna pet the husky bro 
pet it
yes pet the dog
We need image perms uaaaaaaaaaauhghghhh
Lol||l||i||pop|| Art Channel
What did I join to
So, typical Discord server?
no we are very much different ๐
(we are not)
we are (pls help me)
mega slayโจ
XD
it works crazy well
which is surprising given the fact u prob trained it on shitty quality
care to give me an idea on how training works?
search up kohya ss
???
like XL based training
no
LORA is a fantastic and pretty recent way of training a subject using your own images for stable diffusion. Say goodbye to expensive VRAM requirements and hello to this innovative new way of fine-tuning! In this video I will show you how to train a LORA weight using the kohya ss GUI with less than 7GB of VRAM and how you can then use those LORA ...
thats how i started
you get better over time
yup i get the concept
like
he shows exactly how i do it XD
wait
i have embed?
omg i have embed perms
yup u safe
post on civitai babe
it aint gunna be public >:3
fuuuuuuck
you can have it though
IM READY TO DISTRIBUTE ๐๐๐๐๐
too late
Super cute!
No idea how you get the results you do-
yes more furries
good prompts
dont worry i didnt
Well yea. He showed me how to get some of my results
yup pfp looks good
I'm just wondering how he upscales it- I google an upscaling tutorial and well.. it didn't work at all
Please give me 
You forgot to remove the โsignaturesโ in the images
idc about em
lol
it was a pure test for the lora
so watermarks idc for x'3
I need it ๐ญ
Yes, give us this Lora hehe
@wet bramble @wide mesa look #1159496755218501692
bro uploaded!!! How pog of u
hot
im gonna die
sd literally last step
ruining my piece
by mozacing or turning it into a water color or scrunched up paper
idk how to fix
ive tried changing vaes
did nothing
and i cant upload images for some reason
qo moment
fr
Platform?
I need to see this
@rustic raft has reached level 16.
Automatic1111
Stable diffusion webui basically with controlnet extension

u a trusted member now so go all off
yayaayaayayayayayaya







