#🏙│ai-images
1 messages · Page 15 of 1
hmm...
It would be great if it were possible to use text or characters in stable diffusion
make a swedish guy
a mango seed
thats 1 pixel too thicc
top and sides
actually.. wait
ITS NOT EVEN CENTERED
It’s because it’s from ikea
They had to build it themselves
And failed
have u ever read instructions from ikea?
Yes
this man did not read the instructions
Make Mario Teams Up With Sonic Roblox And Minecraft Steve
@normal cradle do ur work
hello!
minecraft steve
what is krubby doin there?
good
promt?
I mean Send your Prompt
i might my guy
yeah just make sure to have the embeddings :3
how do you make an image
Ai
rn im in the process of fixing the dataset keywords for a personal lora im making lol
im halfway through
Make him say "nah I'd win"
lol
Not sure how well the images are going to turn out considering it’s an AI image dataset being turned into more AI images. It carries over the small details and imperfections that the original images and 2x them. But if i’m wrong and it works pretty well then i’d say that’s impressive.
We’ll see what happens 🤷♂️
Just for funsies
I’ll finish fixing the keywords tomorrow though
Yeah, just make sure not to prompt in backgrounds and I think you'll be good
for style training should this include description of the background?
😭
No?
no for character and style lora’s, you never want to prompt a background
ah okok
how do u do ai image thing?
ok thanks
Ayo? @mild raft level 1 !!! 
Hello, please suggest some prompt resources for sketchnote business style in Midjourney. Thanks
Look up PromptHero and see if you can find something there.
Running girl in jungle road
What style
big
@tawny peak Where can I Check the picture to know if it works or not?
(To Check This Picture)
Well.. Check if there is such a thing in Other Photos
Idk how to use controlnet in the UI you use.
No no.. I mean Not in colab but Somewhere on the Internet
Or is it impossible not to?
Maybe with a huggingface space or something, but I haven't checked
what do you need help with? @graceful lark
4 outta 4 for correct hands
i hate waiting for models to train
Don't worry, I'll try to do everything myself 😅
wdym?
never mind
can't you do that with SDXL?
honestly comfyui is the arch linux of stable diffusion
Yeah 100%
Though arch has gotten easier to install so maybe more like the gentoo or some other hard as hell operating system to install
What is this from
?
Mk
@shadow bane If true.. I couldn’t find anything like this anymore. Only this picture was
Finished fixing the dataset, now I am training
good luck
Though I thought you used Bing AI, did you finally get a computer that can handle SD?
This is colab
I'm unsure if my computer can handle SD Training, it may work with regular SD (ComfyUI)
Yeah i was referring to normal SD not training, but yeah it's still requires more horsepower than generating.
Mhm
And yeah you could technically run any UI you want to if you use an LCM model
That's why I create all my Loras with Colab anyway haha
Same here, might as well
Okay it finished but now im tryna find the checkpoint used for whenever i made images previously and goddamn civitai is so slow ;-;
It's most of their new features they keep adding, you can turn most of them off in the settings
And I just somehow hacked together a 3 image dataset into an 11 image dataset.
Where in the settings though
kthx
I'm back
Lora works decently so far, glad it works on the main outfit in thedataset
Mhm
Will you Send Her to the site?
It's a boy
Yes, I know
Maybe idk 
im just testing it for now
No problem
Is this 1 Lora?
Yep
Very Cool for 1 Time
I was using what I was using previously when I was generating SD images
I didn't do it very well 1 Time
Same checkpoint, VAE, easynegative, etc.
From what I understand it's basically something to remove the dull colors of SD images
@~@
And it can enhance it or give it a different feel I think
If the images don't look right, the VAE helps it look better
Then I thought correctly 😅
Yea
Obviously the "realism" portions of the dataset that I have won't necessarily work as well if I have a digital art-esque checkpoint
ex. these types of images
xD
Nice, im surprised
yea me too lol
here's how the realism ones look with the checkpoint i have
hardly looks realistic lol
Cause I'm using an anime checkpoint
Specifically this one
https://civitai.com/models/137193?modelVersionId=151385
Create a Gold Experience with CAT! No VAE needed To fully utilize the Gold Experience please consider listening to Giorno's Theme the first time yo...
Aaaaand I already broke it oops
tried to download very bad image negative and it legit broke i think
ksampler has a pink outline around it
should prob to go #🔍│help-ai-art for this
Girl in jungle road
Ayo? @wheat aurora level 1 !!! 
comfyui isnt hard to install
I know it isn't hard to install, I was just referring to how hard it it to use by listing linux distros.
You cannot generate images from this server you have to use another website or service.
yea
unless ur fluent in it
If i had started using stable diffusion tools later than when I started I would’ve chosen comfy.
Because I has used A1111 for the old NovelAI leak and just kept going after that
:(
ai-image-art a guy
Dude
lol
?
Hello
hii
OK if AI did that that actually amazing
Yes the AI did it
Beautiful
Ayo? @crimson nacelle level 1 !!! 
@wicked moth
thanks timbo
They think they are funny
People don’t like ai art
Edgy moment
I dont even know moment
It’s because it’s trained off peoples art
And people don’t want a robot to take over their career
I use ai when I can't commission te artist in question
my pfp was 20 dollars btw
but for certain artists like ones for games I just use ai since I can't commission em
i cannot help but find amazement in the fact that things like diffusion models even exist. the math behind them is so incredible in how it just works... from image to noise to noise to image, it's fkn mind-blowing that the concept works as well as it does! it's soooo fkn cool
xd I love how the lighting is used here, totally loved it for a cover xd
yess, indeed. We're in the #🏙│ai-images 🥰
same same
hot
Anyone know the best ai for image to image? trying to turn pictures of my friend group into arcade fighter characters on mac.
Is it an intel mac or M series mac
metal diffusion maybe
bro????
M series
damn what model did u use for that
Try out an app called Draw Things in the mac app store, it’s the best program for mac currently and Includes Custom Model support, txt2img, img2img, controlnet, LoRA/LyCORIS, and all the other stuff that other apps have, but it’s built to work the best with mac’s and the new M chips
It’s completely free and you don’t have to sign up for anything as it uses your own hardware to generate
boykisser
Anime
oh yes pikachu i didn't know you were real lol
does anyone know what the ai art app is called
anyone?
?
Ayo? @misty ferry level 1 !!! 
nevermind then
there's multiple
fake
Make Sonic And Mario Arguing But Minecraft Steve Doesn't Like It
always left hand
Works with iPhone too
hard image
awesome image, what model did you use here? I'm very interested in creating something similar
/imagine
is it ai ?
It was made through midjourneyV6,
If you plan on modifying and twiking your picture I would recommend stable diffusion tho
Just a gif
Now Make Sonic And Mario Fighting But Steve Is So Angry
Ayo? @lime stag level 1 !!! 
do it yourself with https://bing.com/create
last one is 8K lol, also have 16K version but too heavy for discord
anyways here's my bunch of best cute ai art, never shared on this server so there's a lotta them xD, there's some animations too, made with AnimDiffusion
I wish animdiffusion worked with the model I made. I think it will, I just don't know how to use it properly
I just did
How'd you do the animations?
AnimDiffusion
Is that an app or an extension? I think the animations are super cool, but I haven't been able to find a tool to do simple animations from existing images.
It's an extension for most of the popular Stable Diffusion WebUIs, though someone could've made a huggingface space or something.
Ayo? @tawny peak level 18 !!! 
Oh cool. I have A1111. Thanks for the help!
Ayo? @chrome ember level 1 !!! 
No problem, it also installs the LCM sampler as well, and it's the recommended sampler to use as you only need 5 - 10 steps to completely generate an image
it also speeds up the process a ton
I love how is GPT-4 with making images of copyrighted characters, it will bend the rules by itself by saying to draw a character looking like X character here, and then describing the character's traits.
Like here is one with Shinji and Asuka that goes so hard.
Bro that does not look anything like shinji though asuka looks normal
Still it’s funny that dalle 3 brags about having all these new features and the best images ever, yet you can’t even see them all because they locked them away
and you need complex prompting to avoid the wall and force them to generate
very true, i like how clever is chatgpt, but sometimes its not perfect
Like i asked to make a basketball movie poster with Graduation Artstyle, and it did this
I just like stable diffusion because when i prompt something, it just works
"a popular 2000s abum" it twisted the prompt
but the colors fit
i am a fellow LORA enjoyer, because some characters I like AI to draw have very intricate details that DALLE messes up
There’s one part of dalle 3 i really like and that it has very good pixel art capabilities and also very good dithering effects for stuff like tvs and old monitors
i agree fully, and its good with realistic stuff
Like the character in my pfp has a distinct chest marking. WIth a DALLE it will forget the detail or misinterpret the detail significantly
I use LORA and it doesnt get the detail perfect, but its way closer than DALLE will, it's almost like a human error.
Sometimes dalle-3 can look really cool and good, and other times it can look like 2022 stable diffusion
2022 stable diffusion moment -_- and openart too
it at least doesn’t go into dalle mini terms of quality
True! I need to get better at prompt engineering AI art
Using stable diffusion interrogate or looking at other peoples prompts on sites like prompthero can help out pretty well
When i draw fictional characters now I always put "a character that looks like character X" instead of just character X's name on DALLE
With LORA I don't need to worry i just need to download the checkpoint and make basic prompts and boom
yep
well thats just a bunch of adorable heaven.
thats just a bunch of minors
wearing revealing clothes
chibi's are better big bulb head i find it cute if your looking at it "that way" thats on you i also like teddybears and pink so if you dont like that hate away.
AnimDiffusion
Through comfyUI
I use a1111 for images but animDiff is way more limitated in a1111, that's why comfyUI is better for this, i can send you a whole tutorial in DM
I would actually love this. ComfyUI ahs been a bit daunting to get started. I got AnimateDiff on A1111 and have been struggling
I'll send it in About 1 or 2 hours, it's huge and i'm not on my pc right now so i can't copy-paste
Take your time. Thanks in advance. I have to get to bed anyway for now
I woke up about 4 hours ago lol, it's 11a.m in France rn
Ayo? @obsidian pagoda level 2 !!! 
It's 2am for me
Damn xD
Bro
You basically already said that before dawg
Also i think you’re underage based on your pfp and username
Now do any of you have ai art of Godzilla
Ayo? @stoic heron level 1 !!! 
2nd one is cute XD
Now do any of you have ai art of horse
#🏙│ai-images Now do any of you have ai art of Godzilla
how to generate?
https://perchance.org/ai-text-to-image-generator I use this but you can use the one from bing to but I have to warn you bing has a lot of restrictions
AI text to image generator. Generate AI art from text, completely free, online, no login or sign-up, no daily credit limits/restrictions/gimmicks, and it's fast. Other AI art generators have annoying daily credit limits and require sign-up, or are slow - this one doesn't. Use this AI to generate high quality art, photos, cartoons, drawings, anim...
ok bro
Ayo? @shadow bane level 1 !!! 
thx btw
np
for some reason they say this is rated R for adults only
i can see why cuz people are stupid but it's pretty normal for
i was very generic anime girl twintails "go"
a
well if you show this to old people they will die on the spot parents think this is not good for their kids and it goess on you just create what you want
Ayo? @stoic heron level 2 !!! 
that is a raw as hell heavy metal cover
now that you say it it is
I need an album cover for my artist Kdiv coco, The song is called "Boys Wayward" . can anyone help?
boys wayward what species you singing about humans demons...
its about a song saying boys are stuuborn,gang song fast cars. its an afroswing song
Ayo? @main basalt level 1 !!! 
Metal asf
i'll see what i can do
not my music type but it's good if i was a fan of that kinda music i guess
great
Is it good enough
So you want something of the GTA kind
Hey ai
?
@main basalt do not post songs with instrumental
YES BRO
Noted
aight i'll try
@main basalt
Ayo? @stoic heron level 3 !!! 
great. can you di it form of an albulm cover
Ayo? @main basalt level 2 !!! 
like this
@icy junco what are you trying to do
@dawn stirrup wie peinlich ong Haha xD
Harry Squatter
We don’t these are real pictures we took

in his asss? damn
how i can use ai for create my pic in this channle
i asked about this a while ago, i use https://bing.com/create
it's nothing like Stable Diffusion but it works really well
I have to warn you bing has a lot of restrictions
i know it does
i feel like that's something you realise the first time you use it
but it's a great first step for people
nah only found out when i used the question anime
i guess
i'm happy to recommend something that helped get me started in playing with ai image generation, but i know that when i started.. i could never have used something like Stable Diffusion straight away
a yes
Apple
🍎
Ayo? @frigid bolt level 3 !!! 
that is adorable 🥰
i know what you are
This will be us
Hgug
Need help?
much more limited though
definitely made for beginners when it says "Negative Prompt (Optional)"
in the UI
Guy named Austria
burger
AlinPRD song Thumbnail called I am Alin
car
Use that website to generate images
Today is Sunday bro
It indeed is
bro how can I upload picture?
Ayo? @blissful dove level 1 !!! 
what do you think?
Cool
How long did it take you to make it
Ayo? @stoic heron level 4 !!! 
Holy shit that's cool lol
gonna try doing a refiner/upscaler type thingy
what did you use ? i have stable diffustion my settings
my model stuff >
suggestions welcome
He probably just used Bing AI
aww
have to be lvl 1 to embed links I think
Yes I indeed used bing for that
فيل ينقز
real.
tried using ultimatesdupscaler again
generated the original image wit anythingv5 and upscaled with counterfeit
you think you're funny for "making fun" of God, and I'm laughing at you for it
???
If you find that making fun of god
You're pick I guess
I can't really change people their mindsets so I'll go with it
@past moth go ahead and destroy this one too
Of course, man, that's playing with God. What are you, Catholic, atheist, what?
And I have to accept the critics of people like you of course
go out naked in the street then
Nah why would I
what do you mean, criticism?
I'm not telling you to become a Christian, I'm saying that I'd like you to respect the Christian religion, you know?
??? what is this opinion lmao
let people make anything they want
doing whatever you want is disrespecting the religion of others bro, i'm not asking you to become a christian i'm just saying to respect the religion of those who are christians
uhhhh i do not honestly think this person is making fun of any religion
in my opinion yes, but ok, think what you want, since you can do "whatever you want"
okie
Okay that's what I wanted to hear
And I accept it
True I wasn't
Thought it was cool but he wants me to respect it so I will
Didn't want to offend anyone
i forgot to change the seed but yay theres a background now
Hah that's even better
why you reposting the server
genius
Is everyone here trying to generate images here over like 50 years old
nah bro
I meant like
The people just putting their prompt here thinking it’ll automatically generate for them
no we post the things that we generated on a other site
but i,m not sure but you can genertate in a other channel
I meant people like this
You can't generate images in the server at all
hand looks a bit iffy
Are you still using SD 1.5?
yeh
Yeah, i've reached the limitations of SD 1.5 and moved on to XL. Though the only downside is that lora training is much harder
btw did you manage to train sdxl lora?
No they seem to not work with an incompatibility with LierLA_Lora models requiring some dependencies currently missing with some UIs and some other stuff, though I may try again training as a basic Lyco model and not one of kohya's special models
I was able to train properly, but the lora just doesn't work
If this one doesn't work out though I'll see if embeddings are worth anything at all
Apparently I have to train with either LierLA or C3Lier as none of the other options work at all, so i'm trying C3 as it's SDXLs version of a lycoris model
@tawny peak what kinda wizard math you talking about here
Ayo? @stoic heron level 5 !!! 
stuff :)
what ai do you even use?
Stable Diffusion XL currently, But I used to use Stable Diffusion 1.5
Basically what I was talking about was the different versions of models that Stable Diffusion XL uses compared to 1.5 and how hard it is to get one to work. I just used their technical names, that's all
okay thanks for trying to make me understand it
No problem
anyway going to bed good night
Good Night! It's only 6 PM for me right now.
Not the place to post off-topic videos, that's reserved for general chat. We used to have an off-topic channel but it's gone.
So @tawny peak I used you're ai program and ask the same thing I did for bing to get my profile picture
This is the difference
What are the VRAM requirements for sdxl?
the silly
i wonder if i should apply another upscaler so it goes to like 4096x4096 or whatever
8gb for proper use
that is cute
What are you using for your negative hand embed?
Ah ok yeah that seems a bit overboard in terms of negative embeddings, but if it works for you it's fine. I usually just went with: (worst quality, low quality:1.4), verybadimagenegative_v1.3, (negative_hand-neg), (text:1.2), (signature:1.2), (watermark:1.2), (extra fingers, deformed hands, polydactyl:1.5), (zombie, sketch, interlocked fingers, comic),
Holy moly
what do you think??
very small image
I Created Stickers Using AI 😄
how do we do the taylor swift ai art? i wanna do that but with lamar jackson a football player for this football group lol
Now thats a good question cause you don't wanna trigger the internet yet again lmao
Looks more Lebron James than Lamar Jackson tbh ^
I was messing around in nvidia canvas
What AI
my first image posted here. Will post more of my works along with CivitAI server and here.
( SD 1.5 )
Time to post my anime characters
i been trying to do this using my OC can you do that ?
welp im a fan its small its cute its squishy its CHIBI
i think the hand detailing thing i did works rlly well
Dm bruh. I wanna make Lamar Jackson and Taylor swift a couple and make a twitter account just for this lol
You mean 𝕏??
You mean X.com
You mean 𝕏??
You mean 𝕏??
You mean Twitter.com???
@quaint tiger look thats ur lora :3
wow it amazing gg
can I have a tutorial in mp ? ^^
No Twitter
You mean 𝕏??
No Twitter!!!
He means Twitter
You mean 𝕏??
Stop it I mean Twitter
You mean 𝕏??
I’m done
Trying to argue with a bot lol
I’m trolling lol
wdym?
Do you want me to show you how to use AI and how to draw?
no thanks you can pass ur way
It’s not difficult for me, but whatever you say
I don't really care so u can just ignore , it a waste of time for u I imagine
I'm sorry if you were offended that I didn't answer you quickly
It's not difficult for me
Better make it Slow so that Everything is clear 😅
I did this without a Graphics Tablet
I drew on my phone
Is this a Lora or something you’re using to get a consistent character
@graceful lark
Yes this is LoRA
I created for a friend
I first made some Images for a Friend (AI Generated)
Next I Created LoRA From These Pictures and Done
@shadow bane
And Now I'm Making Pink Panda
New
if you sent me that i would not be able to tell it was ai
good job 

@quaint tiger / @autumn moat / @tawny peak Where is Channel 2?!
what channel?
channel 2? :3
In the Other AI category, there was a channel where @quaint tiger send his LoRA about Boykisser
SD Resources?
it got moved

Do I understand correctly? I can Spam You via DM Ping?
Okay sorry bad joke
thx ❤️ 🍪
im gonna vote in verified
art is art
(it's ai art dw, we saved this one)
We also need to rename #🔍│help-ai-art to something else so people don't keep asking for RVC shit since they can't read
help ai-art

(sorry traditional artists)
Idk what I try to do
I may also make a bot so people can search for Stable Diffusion models here since CivitAI has an open API
I'd recommend people put a disclaimer on their piece if it is AI btw
Did you just delete the real art channel btw
not delete
just archive
maybe I can bring it back tho, this category lookin bare ahh
gave you view perms until I know what I am doing
Thanks
Ayo? @proud sequoia level 2 !!! 
Current ComfyUI setup, it's good enough for me
so you just gen an image, upscale it and pass it thru again ?
Yeah, also called hires. fix
u can try ultimatesdupscale for that i think it works well
Yo, what does this do?
Make images with AI
I'd like to train a model to do the type of art style I do. Do you know how to do thiss?
I have 30k of photos, and 30k outputs of the art made from the photo.
Yeah, I've trained models before. Is the art style realistic based or more cartoon based?
It's art. Not cartoon but almost there.
Ok, Mind sending one sample image so I can get a basis of what you need to use to train? Here or DM is fine.
I'll dm you.
Ayo? @cloud rock level 1 !!! 
K
ill show mine wen im home
Bro does not know how to use discord stickers properly
A
B
C
D
They Want to Steal Me help
The one who created this Category (I think it was @shadow bane) Thank you!
Ofc
we sortin stuff out
Yeah thanks
I'd say so
Question Why?
it's implying something lewd / nsfw
6 Fingers Very Fashionable
Yeah no negative hand embedding exists for XL yet, or at least not that I know of
But it's still very beautiful, I have nothing against it
I like Blue Hair
The problem is that He thinks I wrote miku but I wrote mikus
Yeah you can usually remove the name from it and it'll still work
Yes but... mikus It's me
everything here is AI
Cat
Yes
Meow
hampter
He tried
Scary
A boy in a black sweater and gray shorts, Age 14 years old, Boy, White Hair, 1boy, solo, dark yellow eyes, full body
these cool withered garchomp anamatronics AI made
eh, not great, but not bad
SD?
Bear thinks it is a dog
Scary when you see a big boy with 7 toes one each side
is it good???
Also I think I finally cracked the code on how to make a depressed Taylor Swift
Well, I'm sorry, I didn't mean it on purpose
Even though in hindsight she isn't really sad in the traditional sense
What program was that generated on anyways if I may ask?
I Use Colab Because I Have a Not Very Good Laptop
Or do you mean LoRA and Model? 😅
What version of Stable is it on?
The colab is both 1.5 and XL
But free tier with xl is slow as hell
Don’t mind me, just compressing my AI images into zip files and clearing the uncompressed ones off my phone so I have more storage
i have like
40K-50K images sitting on my phone unused
and im transferring them to my pc

might put an assortment of them online, labelling them too
(i wonder how google images is gonna deal with all the mass image spam i may do in the future)
not on google drive cause apparently those dont get sent to image scrapers
All from bing ai and google imagen? Damn how many accounts do you have?
idk why, but i can always generate as much i want with bing
i never hit somkinda limit
I have a maximum of 15 Fast Coins
i use their app though
It used to be 100 a day boosts with the dalle-2 version of bing ai
Yes but it's 15 now
@quaint tiger you have 100 Coins Or do you mean that you can create as much as you want but you have to wait about 3-4 Minutes or what?
uh like 8 or 9 on google, prob like 20/30 accounts on bing
i dont even see coins in the app
Oh this is my cart
it just lets me talk to their co-pilot thingy
?
and i can let it do whatever i want o-o
thankfully i can do this easily and somewhat quickly
15
i use the mobile bing ai app
Bing or Edge?
bing o-o
holy hell 
actually this accounts to around like 26.75K
I reached 20k images over an entire year, how old is your first image
i'd have to check
that you’ve saved at least
?
i have to search through a buttload of images
just click date modified when looking through the folder
I mean thats not all weird... I had like 100k images at peak lmao
Damn
Granted this was when Clip was free lmao
and I had to remove the watermark for each one
@tawny peak You also got 15 coins**********? Yes?
WHAT?
But thanks to some open source buddies at "Google"
COINS!! I MEAN COINS!!!!!!!!!
I have free access to Stable XL
AYO
But its like the most basic of Stable XL 😛
Wait how would I search through them anyway
But yes i do
I'm sorry please forgive me
I just started using SDXL and i’m never going back
Though the requirements for training are absurd
all the current images I have go to over a year ago, in this month
i think i still have more on my phone i just havent checked that
Ok
first one on my phone says around january of 2022
For AI images or actual images
lemme check my macbook now
My Gallery Has Only 2 Folders
DCIM ~ What I Photographed
Unkown ~ AI Art
you remember looking glass
its basically dreambooth before dreambooth
Ah
Yeah
Is this SD?
No
ruDALE-E?
ye
What does ru mean?
Oh, are you Russian? Right?
it was translated to english
Difficult
so people could use it
and make checkpoints for the generator
for making certain types of images
Floridaman Language
Did anyone by any chance replicate the graphics of GTA 6 into one of the AI art programs?
man i went wild last year lol
these were some of the first sets of types of images I generated with Bing
(or Dall-E 2.5 as people called it at the time)
Yeah, theres a few
also i loved doing this style when Dall-E 2.5 was available, which was "grainy/dirty polaroid filter"
sometimes at night
new pfp*
i actually made some of these my pfp
this one
there was a variation of it with a black hoodie and a different background i made as well
Found it
Can I think that the White Tiger (Sorry if it's not Him) Is this your Fursona?
Was gonna be that I think but I backed out from it
T-T
anywas look at this transition from 2.5 to 3 lol
why?
got interested in other things
who
I Mean That AI (bing/SD) Can't Draw Shark (Fursona)
o
@void dew this is what I meant when I said shirtless buffed arctic tiger 😭
He sent it and deleted it quickly
I deny the furry allegations.
Dawg it was an accident
I still had it copypasted
I know 😭
No you don't know 
Not sure if this counts tho, spoilering it just in case
This is comedy gold
True
Here’s another version
gato
@tawny peak What magic should I use to make him draw Very Well?
Yes something like that
Because Everyone Else Draws Well But Shark Isn’t Very Good
Im not sure if the UI you're using has interrogate
No interrogate is a tool that collects a prompt from an image
Naw i'm saying you can do it, interrogate isn't necessary
T-T
Okay, I'll ask differently
Wait
What do I need to write so that He Draws a Face Very Well?
Ayo? @graceful lark level 26 !!! 
I wouldn't know
Holy shit why are these so cute
what?
But if it doesn't violate Rules Discord TOS or Server Rules Why not?
i mean that's just me do what you want post a furry thing or 2 your choice i,m just not a fan of furry's that's all
If you don't like it then I'm not sure why Others shouldn't like it 😅
Better forget it because this channel is not needed for that
indeed
@brittle ocean Don't mind the small step count, OneTrainer does stuff a bit differently, but my 3060 with 12GB of vram runs at the same speed as the t4 colab when training. I'll just have to see if this works or not.
and it isn't throttling at all!
ok ok
It’s 26 steps per epoch apparently with the small dataset i’m using
IT WORKED, though I need to change settings.
It also only took like 30 minutes to train


@quaint tiger sweetest thing evr
////<
Yes, I agree with you, he's cute
Here answer
Ayo? @stoic heron level 6 !!! 
cat be driftin like a pro
He does
how did you get it so chibi?
Bing
My new home
Burn
hot
You are recruited to work as a Demon

@shadow bane
oop

