🏙│ai-images

[shadow bane]

yeeeeesh pony diffussion is good

[left carbon]

can someone dm me pony diffusion

[tawny peak]

The whole model or the link to it?

[left carbon]

link to the model

[quaint tiger]

just a lil something i made for @shadow bane

[shadow bane]

:)

:)))

:))))))

[shadow bane]

https://tenor.com/view/team-fortress2-scout-gif-23871356

[shadow bane]

[wide chasm]

walt disney - trump lol

[languid bobcat]

tis a nice car, beware of potholes though

[tawny peak]

One speed bump and that thing is gone

[wheat radish]

if yall wanna test https://discord.com/channels/1159260121998827560/1228715105127235646

[tough forge]

#🏙│ai-images create a image of a digital high definition digital alpha wolf standing over a pray [food]

[cloud basalt]

it doesn't work like that...

this channel is just for people to post their sd stuff

[quaint tiger]

Are we not gonna talk about the name they have?

XD

[bronze kayak]

[shadow bane]

Why do people still do this istgoodness

[cloud basalt]

oops, didn't notice xd

someone already banned him it looks like

[quaint tiger]

Yea i notified litsa

XD

It was funny though imo

Makes for a good screenshot xD

[shadow bane]

Are you gonna delete the message or

[cloud basalt]

nah no need to

[shadow bane]

Oh ok

[uncut slate]

That's me:>

[earnest vector]

Woof~

[errant stump]

heya, when i put a reference image into foocus (let's say of a man) and i want the ai to make the man do a specific action, how do i prompt it?

[leaden oyster]

Ayo? @errant stump level 7 !!!

[errant stump]

just "a man doing x" and all the other stuff or is there a way to specify the ref image

[graceful lark]

OpenPose (ControllNet) + Use #🔍│help-ai-art

[wise shard]

Haze Of Confused

[void dew]

@opaque spear

LMFAo

@shadow bane

[opaque spear]

michel jumpsc

[uncut robin]

The most randomized result I've seen so far lmao

[tawny peak]

phish

[uncut robin]

Yes

[tawny peak]

penguins

[unborn basin]

Trying out meta ai. It's getting better at making images with text prompt.

[tawny peak]

Try out stable cascade, the open source alternative to DALLE-3: https://huggingface.co/spaces/multimodalart/stable-cascade
It'll soon be replaced by SD3 which is looking very promising.

And no watermark using this

[autumn moat]

I wonder how good the 800m will be

And the 6b

Most people just ignore them but I feel like they might be good ngl

[tawny peak]

Yeah same, i just hope the people who have the hardware to train models will utilize them.

[autumn moat]

Yep

[stoic heron]

Wish one of the 3 is better

[stoic heron]

This one count's to

[stoic heron]

A space dog

[quaint tiger]

o-o

[stoic heron]

Question what's that face for?

[stoic heron]

A yes confusel me even more

[stoic heron]

https://tenor.com/view/gun-glock-gun-reload-glock-gif-23914966

[leaden oyster]

Ayo? @stoic heron level 8 !!!

[stoic heron]

Last chance

Explain

[stoic heron]

Aight guess you can't ask a furry a question

Of to kill my self aigen

[quaint tiger]

oof

[stoic heron]

https://tenor.com/view/skeleton-gif-25720883

[quaint tiger]

aight well goodluck with that owo

[bronze kayak]

[shadow bane]

https://tenor.com/view/wheels-on-the-bus-mario-rap-mario-meme-mario-wheels-on-the-bus-wheels-on-the-bus-go-round-and-round-gif-26977128

[stoic heron]

https://tenor.com/view/wall-gif-23473643

[shadow bane]

True

[graceful lark]

[crisp onyx]

Dog

[quaint tiger]

just a random generation while training a new lora

[toxic hollow]

which ai yall recommend the most?

[quaint tiger]

stable diffusion forge

comfy UI if ur tech literate enough

[tawny peak]

lol anythingxl does jitome eyes kinda goofy

[shadow bane]

SD3 sorta fire i guess lol

(at https://sdxlturbo.ai/stable-diffusion3, idk if its the real sd3 or not)

(right is also slightly modified)

[tawny peak]

yeah it is SD3, they just released the api for base and base turbo models, but not the full thing yet

[shadow bane]

ok

[tawny peak]

i'm just hoping system reqs aren't crazy like with stable cascade

[graceful hornet]

Create a Profilepicture witch a shop

[quaint tiger]

you dont generate ai pics here

you post your results

[tawny peak]

he's not even in the server anymore

[quaint tiger]

Oh wtf

My cache wasnt updated

Said for me they were still here

[graceful lark]

😮

[quaint tiger]

Cute

[shadow bane]

okay i figured out how to use pony diffusion xl and now im intrigued on generating more with this

[shadow bane]

all of my current ones are listed here (search for all with the tag i specified):
https://www.deviantart.com/mrrfyai/gallery?q=%23pony_diffusion_xl

[shadow bane]

I mean just look at this guy

(im not sure if this is spam cause i uploaded 2 sets of 6 within hours of each other)

[tawny peak]

Tried out stable cascade (before SD3 officially releases) and the quality at least for anime styled models is about the same as sd 1.5 rendering an image at a low res and upscaling it.

Though inferencing speed is about 4-5x faster than 1.5

[quaint tiger]

Ill stay with the pony model for now

[wicked scroll]

stable diffusion 3

[pine torrent]

Hi

[sand kelp]

https://www.youtube.com/watch?v=a93ZS2IQHNA

[leaden oyster]

Ayo? @sand kelp level 1 !!!

[sand kelp]

I guess art?

[teal stag]

How did you make that?

[maiden rune]

Marwany through -Nivi V6 on midjourney sever, you could go on a third party software tho, even tho I don't know how

[leaden oyster]

Ayo? @maiden rune level 2 !!!

[lost thistle]

is there a way to chibi(ize) a premade chr i have done so i can get some my chr but like emotes id like to do it off line so its free i have stable diffusion ui offline lemme know i'll be afk a bit just ping me

[graceful lark]

??????

wdym?

[graceful lark]

sowwy I'm very stupid, can you write in other words?:

[lost thistle]

i want to take a chr i own (i have drawn with cgi) and turn it into a chibi with different pose for emotes

possible?

[graceful lark]

Oh yeah

You can create LoRA and try sd 1.5

Or I can try if you don't mind

[lost thistle]

idc im not really good at this im just curious im just trying to make some twitch emotes of my own chr vs having them drawn as i cant afford it xD

[graceful lark]

oooh

You don't use SD, right?

[lost thistle]

sd?

[graceful lark]

stable diffusion

[lost thistle]

oh i have it im just not great with it my actual chr was made with another program though

[graceful lark]

If you don't mind, let me write to you dm Because this channel is not created for this

[lost thistle]

sure i'll pm ya

[modern mist]

i want a picture with ai of dog with human body

[shadow bane]

You can make it yourself, try https://bing.com/images/create

[surreal merlin]

Princess Mindy from the Spongebob Movie!

[shadow bane]

75% close

But aight

[stoic rock]

Wow

[quaint tiger]

very nice, bing ai?

[vapid osprey]

Yeah

And upscaled it locally

[quaint tiger]

nice nice

[sonic sun]

[onyx rover]

https://tenor.com/h8qxTy9quRC.gif

[shadow bane]

Are you a child?

Just wondering

[quaint tiger]

most likely

[long yarrow]

#🏙│ai-images

[shadow bane]

wha-

---

---

anyways (not asking this guy) i wanna make a pony diffusion xl lora for a pokemon using free colab and idk how

Perhaps Yume can help whenever I wake up in a while from now cause I’m gonna sleep soon

[brittle ocean]

#1159496755218501692 i made tutorial

[shadow bane]

Ok

[brittle ocean]

although it looks good from a distance

[tawny peak]

Stable video diffusion can animate existing images but it can also be linked up to generate an image then animate that image.

[tawny peak]

how about now

[brittle ocean]

howw

[tawny peak]

Pixel art loras recommend you take the image and downscale it 4x then re upscale it back up to original resolution using nearest neighbor

[brittle ocean]

so i render it normally then i downscale it and upscale it right?

[tawny peak]

yep

aseprite / libresprite is what i use, though it can be done within the workflow, it just requires a bit more work

[brittle ocean]

alr

[brittle ocean]

wait what is nearest neighbor ?

[tawny peak]

an upscaling method, if you know what bilinear is, it's similar

[brittle ocean]

oh in sd this is nearest upscaler ?

[tawny peak]

Yeah

[brittle ocean]

ok ty

[brittle ocean]

one more question how do u unblur it ?

mine looks blurry af

[tawny peak]

I didn't need to unblur mine, possibly a weird quirk with the upscaler used in sd

[brittle ocean]

[quaint tiger]

my result

[tawny peak]

i guess also make sure each image is png

[brittle ocean]

its working now

[tawny peak]

What setting did you change

[quaint tiger]

im guessing it downscaled using bicubic

then upscaled with nearest neighbour

[brittle ocean]

program i didnt know how to import image to aseprite so i used krita

later I found out how to import images

in aseprite

[tawny peak]

ah ok

[graceful lark]

😮

[shadow bane]

Blaziken

[leaden oyster]

Ayo? @slate ember level 1 !!!

[shadow bane]

???

What’d they say

[dusty kestrel]

If you mean what that user said, they wrote a bunch of unrelated, rather sus things but not related to you afaik

[shadow bane]

‘Kay Ty

[languid bobcat]

Anyone know if comfyui is blocked by colab? We could technically use that for rvc inference.

[tawny peak]

yes, any webui based ai app is blocked of course just for free tier

[languid bobcat]

Improved rvc music video workflow (without hires upscaling):

[graceful lark]

😮

[shadow bane]

Real

[stoic heron]

bandicoot on extreme level

[maiden rune]

🤌

[shadow bane]

Are you sure that’s AI?
Edit: nvm it is

[humble otter]

why post this man

literal fetish art

[humble otter]

really good minus the extra random finger

[shadow bane]

Yeah

[surreal merlin]

Why does it matter?

[humble otter]

because there are minors here…? who also shouldn’t be exposed to fetishes⁉️⁉️⁉️⁉️

people get their rocks off to that, which arguably means it shouldn’t be in a SFW server, no?

[surreal merlin]

I'm pretty sure this is an 18+ server dude.
Plus, what would a minor even know about AI anyway

[humble otter]

definitely not an 18+ server, go into any channel and you’ll find people under 18

i’m 16 and i know a decent amount

considering you’re using a free ai hosting site i’d doubt you know how to run something like stable diffusion locally

if it was an 18+ server there’d be some sort of verification and definitely not 300k+ members

[tawny peak]

Hate to be "That Guy" but rule 1 does say no NSFW, even if it isn't entirely NSFW, if it's enough to make people uncomfortable then don't send it.

Sorry in advance I am just doing my job as a helper :/

[humble otter]

at the end of the day latex is a fetish and fetishes are nsfw

i personally wouldn’t feel bad

[surreal merlin]

Bruh, that's not fair.

[tawny peak]

Sorry, didn't make the rules

[humble otter]

it’s an ai generated image dude get over it

read the rules and apply logic brodie

[surreal merlin]

If it's an AI Image, I should be allowed to generate whatever the hell I want.

[humble otter]

yeah go ahead and generate it, just put it in your stash and not a server with children

you broke the rules, and the only punishment was a deleted image, stop complaining

[shadow bane]

Hello I am minor that knows about AI

[humble otter]

hello i am three minors in a trench coat that knows about ai

[shadow bane]

lol

[autumn glacier]

hello i am a minor that knows about ai

[quaint tiger]

pretty sure theres more minors in here than you realize lol.

[harsh hornet]

Hello I am minor that knows about AI

[autumn glacier]

this guy got cooked help 😭😭😭😭😭🙏🙏

[brittle ocean]

some minors know more about ai than u

[cloud basalt]

we're 13+ brother

[shadow bane]

^

[quaint tiger]

Buddy is underestimating minors

Theres a whole bunch of em that know so much more than i do

And im fucking 30

[autumn glacier]

well tbh newer generations always learn new technology better

(NOT the case for gen alpha)

[quaint tiger]

Its cause minors their brains develop faster, and being exposed to the technology available to you, you simply learn it easier

That same will happen to yall when you reach 30 years

Then those minors will know more x'3

[autumn glacier]

yea makes sense

[clear pivot]

Hayato Tani from Takeshi's Castle

[quaint tiger]

ahahahah

nice XD

that outfit is so recognizable

[clear pivot]

Thanks

[humble otter]

im very proud

[sonic sun]

:

[clear pivot]

Hey, Tani

[arctic gyro]

Heh... Why my prompt is unsafe??

Can't i put a tank there?

[arctic gyro]

Can i have an anime prompt? :)

[leaden oyster]

Ayo? @arctic gyro level 2 !!!

[arctic gyro]

Oh nevermind i can

[shadow bane]

What sites do you guys use for high quality art?

[graceful lark]

colab, bing ai

@shadow bane #🔍│help-ai-art message

[arctic gyro]

Is it possible to make ai art with more lines in bing?

[shadow bane]

Use copilot

[cosmic sun]

RONALDO

[shadow bane]

sui

[clear pivot]

Ponyfied Hilary Flail.

Hilary Flail too

[shadow bane]

Recently watched episode 1 of Star Wars in theaters

[eternal silo]

A fellow Inazuma Eleven fan, I see.

[clear pivot]

Her

[rancid cave]

furi

[normal cradle]

Him

[graceful lark]

😮

[austere garnet]

Twilight Sparkle

[tiny warren]

love

[shadow bane]

I wanna do this myself too >m<

[graceful lark]

I used LoRA some stickers (https://civitai.com/models/109431?modelVersionId=117923)

[shadow bane]

Ok

[sturdy violet]

The Tianamen Square

[minor bough]

Gay love

Love

[graceful lark]

@shadow bane 😮

[shadow bane]

Crazy

[graceful lark]

mb

[tiny warren]

abuse

[tiny warren]

the original sin

[arctic gyro]

I can't change the image size?

Using bing

[fathom fox]

Lion and rabbit

[graceful lark]

Yes, if you use bing then Size is only 1

[leaden oyster]

Ayo? @finite warren level 1 !!!

[graceful lark]

@finite warren What do you want to do?

[finite warren]

black cat

[graceful lark]

Private Sticker

[finite warren]

its black??

[graceful lark]

Use bing for this https://bing.com/create

[finite warren]

how to make it

[graceful lark]

=-=

Open this site and then you will understand what to do

[shadow bane]

out of context

dawg

you just copied my message

[finite warren]

thanks

[finite warren]

thanks you

[leaden oyster]

Ayo? @finite warren level 2 !!!

[graceful lark]

np

[tawny peak]

Chill with the message spam, it obviously didn't work so don't keep trying it.

[zinc dragon]

e

[shadow bane]

Bro why do people say very few things then leave 😭

They literally said this and left

[tiny warren]

https://creator.nightcafe.studio/creation/RwiaUOmByLAL4cygPIc4?ru=veritas16

[tiny warren]

Hope as a path to salvation

[tawny peak]

They figure out they can't generate images, then promptly leave.

[shadow bane]

🤷‍♂️

[shadow bane]

Rapper Mewtwo

[shadow bane]

@shadow bane what type of SD u use? SD XL or other SD?

(stable diffusion)

[quaint tiger]

i think he uses bing ai?

[shadow bane]

^

[lucid urchin]

https://youtube.com/shorts/xv6xj2wPCP0?si=0wygqsh08HymANt4

[tiny warren]

the artist and the muse

[languid bobcat]

not really art related but I'm almost finished the RVC inference nodes in comfyui:

[shadow bane]

That’s crazy

[cloud basalt]

holy shit

[cloud wigeon]

steve doing slipback

[warm ibex]

Ho lee shit

[shadow bane]

These look raw as hell

[bronze kayak]

third one is scary

[shadow bane]

Yeah

[languid bobcat]

one of my failed creations

is the blurriness something that supir/topaz/oss upscaler can fix? @tawny peak

[tawny peak]

you mean for the background or the whole thing

[languid bobcat]

the foreground, I think the background is supposed to be blurry (out of focus)

[tawny peak]

yeah i'd just try upscaling each frame then converting into a vid file

Are you using just base animate diff or hotshot xl?

[languid bobcat]

pure custom sd1.5 model+animatediff+hyper lora. no fancy ipadapters, controlnets, or qr monsters 🤣

[tawny peak]

hotshot xl is just base animatediff but using a different motion model and it's of course SDXL

The gen times I posted were on hotshot xl

not 1.5 animatediff

[obsidian ravine]

I just created a pixelated cat lora

[tawny peak]

Downscaling and Upscaling again didn't work for this, it required a whole redraw

[languid bobcat]

I only have 8gb vram and 16 GB Ram 🥺

[tawny peak]

SDXL can run in only 4GB

[languid bobcat]

I get oom when combining a 4 min image batches at the end at 768 resolution. Wish the video combine node was more efficient

Do you mind testing my workflow with your XL after I release it?

[tawny peak]

sure whenever i get time to i will

[tiny warren]

breaking the spell

[leaden oyster]

Ayo? @tiny warren level 2 !!!

[shadow bane]

😋

[obsidian ravine]

[shadow bane]

bro that looks cursed

[languid bobcat]

these are some interesting audio generated key frames

[shadow bane]

Ghibli inspired walk in the Welsh countryside, using Udio and Midjourney
https://youtu.be/LokgSHlN3tc

[arctic gyro]

Is there any site besides bing makes good ai art?

Bing one blocks my prompt

I was just trying to make a character sleeping

[stoic heron]

https://perchance.org/ai-text-to-image-generator I use this one sometimes

And civit.ai is also a site you can get models from characters and other stuff

[graceful lark]

Online (Website):

• Civitai
• Frosting
• Leonardo AI
• SeaArt
• Yodayo
• Tungsten
• PixAI
• ImgnAI
• Perchance
• NovelAI
• Bing (only SFW)
• WebSite?
• Craiyon
• Dezgo
• Meta ai
• Tensor
• Ideogram

Online 2? (NO GPU):

• HF | cagliostrolab/animagine-xl
• HF | stablediffusionapi/pony-diffusion-v6-xl
• HF | dataautogpt3/Proteus-RunDiffusion
• HF | multimodalart/stable-cascade
• HF | prodia/fast-stable-diffusion)
• Colab | Stable Diffusion 1.5 (support pony)

Local (GPU):

• Github | ComfyUI
• Github | Stable Diffusion Forge
• Github | AUTOMATIC111/Stable Diffusion

Phone:

• Android ~ #1196292814875672707
• Use any apps at your own risk

Helping:

• Use: #🔍│help-ai-art

These are all the sites that I tried to use

[fluid escarp]

Holaaa

Quiero un imagen ia

[graceful lark]

Here are just useful links, there are also SD and others

• https://huggingface.co/spaces/r3gm/DiffuseCraft
• https://huggingface.co/spaces/John6666/DiffuseCraftMod
• https://huggingface.co/spaces/TheStinger/Ilaria_Upscaler

[quaint tiger]

puedes generar imágenes usando: bing ai / stable diffusion (probablemente haya más opciones, pero esas son las más fáciles de hacer)

[graceful lark]

I like it!

what did you use?

[neat nexus]

Pony Diffusion V6

[graceful lark]

cry

I am upset because of you

[shadow bane]

Why

It looks good

[tawny peak]

I've told you many times now, you can use it on the Colab.

[graceful lark]

yep

But I Sent the sites (and colab as well) So that Others would know Because, as I was told, colab is a difficult thing for Others

Although I think not? 😅

aah

I understand what you mean

Yes I'm stupid sowwy

[tawny peak]

For some people it's pretty straightforward, for others not so much

[graceful lark]

Honestly speaking colab PONY Even more difficult even for me

Because 1. No LoRA This is the first minute
2. no img2img This is the second minus
3. Colab doesn’t always work (I mean that the GPU even When I didn’t use colab doesn’t want to work, only pony others work) This is the third minus

[tawny peak]

Here i'll try the colab with pony xl real quick, ill show that it works

[graceful lark]

@~@

I'm too dumb for this please do it

[tawny peak]

It works!

[graceful lark]

Which colab are you using?

[tawny peak]

The same one you use

[graceful lark]

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

A?

https://colab.research.google.com/drive/10Naa1SiIy0CA7bjk0q1rCcMMza6aWXpy?usp=sharing

?

[tawny peak]

I can't access that, you have to send the original colab link not a drive link

https://colab.research.google.com/github/R3gm/SD_diffusers_interactive/blob/main/Stable_diffusion_interactive_notebook.ipynb

This one?

[graceful lark]

wait.. are you using pony in this sd?!

[tawny peak]

Yeah

[graceful lark]

=-=

It's a joke?

[tawny peak]

No

[graceful lark]

Because I can't even use yiffmix 36 What pony?!
If I create through this colab my image is completely noisy

[graceful lark]

I'm dumb for this T-T

[tawny peak]

You need the SDXL VAE, the one it auto installs is for 1.5 models

Your model page needs to look like this

[graceful lark]

Hmm.........

Let me try (btw if I use pony I can't use my old LoRA?)

[tawny peak]

No old loras don't work. but you can always train on the SDXL colab and just use pony xl

[graceful lark]

yeah otay

[tawny peak]

as the training model

[graceful lark]

• Type 2 LoRA doesn't work (Or I'm really too stupid)

Lacosia?

idk wait

LYCORIS

I wanted to try this but it doesn't work https://civitai.com/models/74596?modelVersionId=79327

[tawny peak]

No, they have to be SDXL and SDXL only, SDXL Lycoris, and Loras will work

[graceful lark]

Try using this

[tawny peak]

Also for 1.5 models theres no Lycoris Support only base lora

[graceful lark]

oh this explains everything

stablediffusionapi/pony-diffusion-v6-xl?

[tawny peak]

Yep

[tiny warren]

Cupid's First Strike

[tawny peak]

Make sure you do this as it is 100% required, set sampler to Euler A, Resolution to 1024 x 1024, Then add score_9 to the beginning of your prompt

[graceful lark]

otay

pwease wait

I have too many questions...

Why didn't it work when I tried to do this before?!

[tawny peak]

You didn't have the SDXL Vae, it's what controls the finalizing of the generation

[graceful lark]

not NSFW!!

[tawny peak]

Just be careful with sensitive images, some people think differently about images than you could.

[quaint tiger]

Poyo finally able to use pony xD

[tawny peak]

I love this character I prompted up a while ago

[leaden oyster]

Ayo? @tawny peak level 30 !!!

[graceful lark]

@~@

[graceful lark]

i really like it ❤️

[autumn glacier]

mia lora for sdxl soon :3

[graceful lark]

😮

[steel moat]

1 attempt, after 3,000 thousand generated photos, I can finally say that I am able to generate what I want in 1 attempt

[graceful lark]

cool

[steel moat]

https://media.discordapp.net/attachments/922954276643426325/1239550814662561832/Default_its_night_on_a_pirate_ship_with_one_piece_calm_sea_mon_0.png?ex=66435527&is=664203a7&hm=6627968fa17247e07ebd51cbc9eb4a10080503a24779d2835e1eade178e6363b&=&format=webp&quality=lossless&width=687&height=390

[graceful lark]

😮

[autumn glacier]

either i figureo ut how to do it myself or yume will just train one for me .w.

[graceful lark]

You can use https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer_XL.ipynb#scrollTo=OglZzI_ujZq-

[autumn glacier]

nah i have a server for using everyuthign

rtx 3090 and 60 gb ram :3

[graceful lark]

smile

In any case, I'll paste this here

[graceful lark]

tail...

[serene storm]

maid poyo ??

[graceful lark]

😮

[serene storm]

bruh

do they know this is not midjourney

[graceful lark]

oh wow..

i`m a bot :3

jk

@warm cliff use: #🏙│ai-images message

[warm cliff]

: /

[graceful lark]

I'm not sure if there is an AI here that draws here on the server

[warm cliff]

I thought it's can...
My mistake

[leaden oyster]

Ayo? @warm cliff level 1 !!!

[serene storm]

nah its fine

[serene storm]

i think gpt can(?)

[warm cliff]

: )

[graceful lark]

I'm not sure about this, but there really was a gpt here that draws, but now I don’t know if it exists Because I don’t use Discord often because I have exams soon

Let me check this

mmm nope gpt can't draw #🤖│bots message

[serene storm]

guess it cant anymore

[graceful lark]

Yes, he used to be able to do it, but now it seems not

sowwy

[serene storm]

eh Bing works just fine

[graceful lark]

Yes you can use bing or the other site I wrote above

[sharp geode]

How can I create ai pictures?

From text to image

??

[languid bobcat]

You can now create ai pictures from audio to image 😉

[sharp geode]

How?

My wifi is so slow😭

[graceful lark]

@sharp geode use

[shadow bane]

Doing AI dark fantasy art style after hours of the surgery i had in the hospital (i'm on the hospital until tommorow lol)

The AI did great, ngl

[shadow bane]

Also i used DALL-E 3 for this if u ask

[pine yew]

how do i ask for a picture ?

[graceful lark]

#🏙│ai-images message

[tiny warren]

The Nature of Time and the Universe

[graceful lark]

chicken..

https://tenor.com/view/куриная-ножка-и-наггетс-поют-gif-12130326630185892344

[spring jungle]

meowscarada

[tawny peak]

60 sec inference for SDXL on iPhone using SDXL Lightning 4-Step, pretty good.

Supports Non-Pony SDXL Loras as well.

[shadow bane]

More Dall-E 3 gens

[shadow bane]

Second image seems sus 🤨

[quaint tiger]

And a masculine dog aint sus?

Furries in general are pretty sus xD

[bronze kayak]

nice

[shadow bane]

Depends on what they do and how they act individually

[shadow bane]

To me I’m fine with it

[valid fossil]

Who have pictures ai celebrates?

[leaden oyster]

Ayo? @valid fossil level 1 !!!

[valid fossil]

?

[stoic heron]

does Jesus count

Don't know why that ping happened

[magic radish]

I saw that tag.

lol

[valid fossil]

What???

[valid fossil]

Like my av

[valid fossil]

Maybe lana del rey, taylor swift, tom holland and others

[magic radish]

Are you going to model after a famous person? Use a face transfer colab.

"There are popular free tools that work with Stable Diffusion."

https://github.com/lllyasviel/Fooocus

"You can use this to create images using the face of a celebrity."

[stoic heron]

Indeed was a accident

[stoic heron]

Nah they boring Jesus is a celebrity to

[magic radish]

Do you think using AI like Jesus is appropriate? I saw an AI mimicking Jesus on the Twitch streaming platform last time.

[stoic heron]

isai?

you speak with magic words to me i shall not understand

[magic radish]

"Small translation error, I meant to say 'Jesus' but the AI translated it as 'ia owner'. I was referring to Jesus there."

[stoic heron]

A

Still don't know what ia means though

[magic radish]

gpt has gone dumb

edited the text

[stoic heron]

nah

what's big problem with it

he is a man

i do not wish to harm anyone with it

[magic radish]

what happens if he gives false information

[stoic heron]

eeeeeeh

i do believe people will be smart enough to know it's wrongµ

[magic radish]

After all, it's a religious thing, isn't it?

[stoic heron]

yes

[magic radish]

Isn't that a sin?

[stoic heron]

if your religious depends on what you do

[magic radish]

In my religion there is Jesus, but in a very different way.

[stoic heron]

this is one i made

i never seen problem there was one guy before that thought i was making fun of him

[stoic heron]

what way then

[shadow bane]

isa?

[magic radish]

I am not a Christian

[stoic heron]

a protest christian

[shadow bane]

are you a muslim

muslims believe jesus existed but believe he was a great prophet and messenger not the son of god

hes called isa

[magic radish]

true

[shadow bane]

so you are muslim?

[magic radish]

maybe

[stoic heron]

well that is a yes

so what will you do if i start making ai art of Mohamed

[magic radish]

I don't know how correct it is to do the conversation because I don't know much about Islam, but I am a Muslim.

I mean, I'm supposedly a Muslim, but I don't worship.

[stoic heron]

so it's your religion but you don't know shit about it

[stoic heron]

then your not

simple if you no believe or worship you shall not be a worshipper

[magic radish]

I am Muslim but I do not worship

[magic radish]

I believe but I don't practice

[stoic heron]

wrote it wrong

@magic radish have it your way i,m leaving now

[magic radish]

Most young people in Turkiye believe but do not practice

[magic radish]

As far as I know, they didn't show Muhammad's face in the movies, so it could be a problem.

Only if you wear your long hair without showing your face.

maybe it's okay

[shadow bane]

disrespectful but nobody can stop you ig

different to generating an image of jesus

making images of mohammed is forbidden in islam

(im an atheist btw i just have knowledge on religions)

[magic radish]

may be true

no information

[magic radish]

https://youtu.be/lFohVtbnaNM?si=SqlkqppzUkcOsW-C&t=344

meanwhile here he tries, but he denies it

says it's not suitable

then it's true

[stoic heron]

what does it say

[stoic heron]

almost everthing is forbidden

[magic radish]

It is unacceptable to make a picture of Muhammad because he did not want his image to be worshipped

2. none of them can be Muhammad because there were no statues or paintings of Muhammad at that time.

[quaint tiger]

Thats a lot of text in an ai art channel >u> hello there peeps xDD

[magic radish]

No one wants any religion to be disrespected.

[magic radish]

That's why I asked that question.

[stoic heron]

religion nonsense

[quaint tiger]

Im sorry but honestly i dont think this is the place for religion talk

[stoic heron]

facts

[quaint tiger]

Ai is gonna ai

It doesnt matter what it is

[magic radish]

Are you interested in, uh, ia painting?

[quaint tiger]

I do ai art yes

I have that 4090 for a reason hahaha

Then again. I specialize in furry ai art

[magic radish]

What should I use as a tool to make a one-to-one with the face of a famous broadcaster?

[quaint tiger]

Oh you want to make ai art of famous people?

Honestly you would be looking at stable diffusion

[magic radish]

Yeah, I'll put them in cool pictures.

[quaint tiger]

This has multiple different uis

Comfy ui

A1111

SD Forge

Easy diffusion

The most common 2 used currently are SD forge and comfy ui

[magic radish]

https://github.com/lllyasviel/Fooocus I used to use this one but it barely resembles it

[quaint tiger]

I have not heard of it

[magic radish]

it doesn't look like much

[quaint tiger]

So i cant recommend it

Personally i use SD Forge

You can download different models from Civit.AI

People make lora models of famous artists

[magic radish]

I heard there's not many celebrities there.

[quaint tiger]

Have you tried using bing ai?

[magic radish]

Is there a colab that makes lora?

[quaint tiger]

Bing ai is its own thing

Its closed source

As it uses dall-e3 with gpt4

[magic radish]

the reply was wrong

I know Bing ai

[quaint tiger]

For stable diffusion

You can train your own models

You grab like 10 to 25 images of a person

And use it to train your own model to be able to create them

[magic radish]

Can I do it with colab?

my vram is bad

[quaint tiger]

I think so, i myself dont use collab

But poyo does!

Poyo uses collab to train and generate

@graceful lark

You can help this person better than me i think, as i dont use collab

@tawny peak you know quite alot about collabs too right?

@magic radish it seems they are currently not here, but i think they are the best people that can help you when it comes to collab training/generating

I myself dont use collab, so idk how it all works

[autumn glacier]

https://cdn.discordapp.com/attachments/1163146739872698569/1240011301116313620/00001-1189811415.png?ex=66450204&is=6643b084&hm=205d96e8de98e6be616d078935fc0f6e02e7dd8855f06a5a075a1d4ba42acab8&

mia is now real on sdxl :3

anothr

[tawny peak]

You looking to train or use stable diffusion?

my bad i didn’t read, here’s a colab to train loras: https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer.ipynb

[graceful lark]

What?

[bronze kayak]

poyo

[graceful lark]

a?

what?

How can I help you? 😢

[bronze kayak]

hi

[tiny warren]

The Boy With No Choice

[normal cradle]

Ps3 looking ahh background

[autumn glacier]

https://media.discordapp.net/attachments/1205908616192856097/1240028763039858759/00006-1793886457.png?ex=66451247&is=6643c0c7&hm=1123de443e633529bb7cefd2068f1b056b0bfc63a4af50cc34890e0deacf8ce9& the silly

[stiff ingot]

Hey everyone hoy can I use this ai?

[tawny peak]

Gotta either use https://bing.com/create, https://civitai.com or just download a webui and run one locally

[mossy pecan]

Guys have anyone tried Tiktok's new ai yet.

Just tried some, seems very strong to me... Convenient to use compare to SD and MJ.

Here is the platform https://www.capcut.com/ai-tool/home

[shadow bane]

Says it’s not available in my region

[mossy pecan]

Dear you are from?

[shadow bane]

US

[mossy pecan]

Seems not online yet...Do you have VPN? Could get to singapore or Asian region

[shadow bane]

I was just checking if it was available so im not gonna bother using one

[quaint tiger]

Lol capcut makes ai now? Yikes, i have never liked capcut xD

[slow geode]

soooo?

[graceful lark]

4...

030

[autumn glacier]

mm?

[shadow bane]

mm.

[graceful lark]

4 fingers..

[wheat radish]

https://tenor.com/view/snoop-dogg-drop-it-like-its-hot-4number-four-fingers-calvin-broadus-jr-gif-26215289

[tawny peak]

Near perfect modern anime style done only with prompting and a character lora (no style lora)

[autumn glacier]

oh lmao .w.

the lora i used for this is a bit scuffed so

[quaint tiger]

The neck seems kinda long o3o

[obtuse coral]

Thanos`s Pet

made with cringe ai

[quaint tiger]

is that an akita?

it looks a lil different

it looks like some weird crossover between an akita and a retriever

[languid bobcat]

now get her animated

[tawny peak]

Hotshot XL is good at retaining style information so i’ll see if i can.

[tawny peak]

Forgot to add the sharpen nodes but look, it retained the style perfectly

[tawny peak]

Another cool thing i can do. ASCII animation

[fading brook]

Here's your generated image!

inductivegrub

"Big Old Manatee playing the banjo"

Generated by Weights.gg

[shadow bane]

https://tenor.com/view/epic-bot-command-fail-laughs-at-this-user-eats-cereal-gif-23218562

[stable elk]

Bob Ross

[unborn karma]

https://tenor.com/bGwWf.gif

[shadow bane]

If you wanna generate images use one of these: #🏙│ai-images message

[shadow bane]

https://tenor.com/view/jesus-ballin-mars-bars-gif-19910027

[quaint tiger]

was doing some style testing :P

[neat nexus]

Prompt?

[shadow bane]

Its a custom Lora with a prompt

[mystic lotus]

fkkd ┬─┬ノ( º _ ºノ)

[quaint tiger]

score_9,score_8_up,score_7_up,score_6_up,score_5_up,score_4_up, solo anthro, YutaOC, upper body, male, (pink eyes), pink sweater, black turtleneck, ((round glasses, black glasses)), white hair, EARS, dotted brows, short hair, small nose, (wink, one eye closed, :3, blep, tiny tongue, tongue out), <lora:YutaOC SDXL-000005:1>, solid color background
--------------------------
deformed, mutated, ugly, disfigured, double tail, bad hand, bad anatomy, missing thumb, nfsw, nipples, long neck, lowres, big mouth, multiple straps, blurry, dof

[shadow bane]

Not for safe work

[tawny peak]

The bottom one is the negative prompt

[shadow bane]

yeah i know, they just put nfsw instead of nsfw in the negative prompt

[tawny peak]

oh i didn’t even notice that

[shadow bane]

yea

[tiny warren]

The "Crime and Punishment" of a True Artist in Tumultuous Times

[languid bobcat]

I thought it stands for No Fu***** Safe Word

[normal cradle]

https://cdn.discordapp.com/emojis/1165636134538326096.webp?size=48

[humble otter]

kill

[serene storm]

blud left for a week and came back

bro cannot live without ai hub

[humble otter]

i was lost without it

its like my north star

[magic radish]

https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer.ipynb

Is there a colab that I can test that I can also use as a training xl from here?

[tawny peak]

https://colab.research.google.com/github/hollowstrawberry/kohya-colab/blob/main/Lora_Trainer_XL.ipynb it was right at the top of the 1.5 colab

[magic radish]

I've already trained from there.
I'll test it now, most colabs are banned.

need colab to try

[tawny peak]

Works if you have a smaller dataset

[magic radish]

No, the training is over.

[tawny peak]

What?

[magic radish]

I need a colab to try these safetensors

[tawny peak]

Oh, ok: https://colab.research.google.com/github/R3gm/SD_diffusers_interactive/blob/main/Stable_diffusion_interactive_notebook.ipynb

[magic radish]

for xl

[tawny peak]

Yeah the colab works with 1.5 and xl models

[magic radish]

Okay, if I type the name of the model, it will produce it, or should I just type “man”?

[tawny peak]

Take one of the prompts from your dataset, and use that.

[magic radish]

https://huggingface.co/stablediffusionapi/realism-engine-sdxl-v30 I trained using this.
and where to paste the model dirive link.
Both of them.

now I got an error

[tawny peak]

Download the lora model and drop it into the colabs virtual files under the lora folder, re-install the pre-reqs and models again and it should show up under the lora section in the ui

[magic radish]

paste the dirive link and it comes here

[leaden oyster]

Ayo? @magic radish level 26 !!!

[tawny peak]

True that works, but you’re trying to use an xl model on a 1.5 lora, that won’t work

[magic radish]

https://huggingface.co/stablediffusionapi/realism-engine-sdxl-v30

Isn't this an xl?

[tawny peak]

Yes that’s an xl model, the lora you trained and are trying to use is a 1.5 model and are incompatible with xl models

Find a 1.5 model on huggingface or civitai and it’ll work

or you could retrain your model as an xl model and it’ll work with the model you just posted

[magic radish]

I trained in xl.

How come it's not xl?

[tawny peak]

You sent this link, this one wasn’t the xl trainer

I had sent that one thinking you wanted an xl trainer

[magic radish]

but I'm already training from there.

I found and trained myself from there before you told me.

[tawny peak]

Yeah that one still works just fine, instead of trying to use an xl model on the lora, just find a 1.5 model and use it there, the lora will work just not with this model here

[magic radish]

Is there a video?

it doesn't matter what language

[brittle ocean]

https://discord.com/channels/1159260121998827560/1183928977681825883

read

it

[ember jewel]

I found a cool AI that is very simple to use. Using the AI i use for design is incredibly intuitive and straightforward, it feels like it understands my creative vision and brings it to life with ease. if you want i can show a few images of how highly intelligent this AI is. just give me a prompt and it only takes a few seconds.

[tawny peak]

Is it https://bing.com/create ?

[ember jewel]

yeah. its copilot which is the same thing.

[shadow bane]

And he’s gone

[shadow bane]

https://tenor.com/view/but-heres-the-sniffer-cat-sniff-gif-26885257

[agile field]

can someone please tell me about the import settings needed for any Pony Diffusion XL thing from civitai onto the Draw Things app? stuff like "fine tuned image size", "model objective", etc

im at my wit's end with trying to make them work & turn out right

[graceful lark]

why?!
(pony 15 eph)

sd 1.5

T-T

Why does this look better?! (1.5)

[graceful lark]

so.. I just realized that I forgot to add the main texts for this..

aa

T-T

[brittle ocean]

sdxl checkpoint issue or change weight of lora

[graceful lark]

||(Now there will be a very funny joke for @quaint tiger)||
My LoRA has 58mb, 15 eph

[graceful lark]

sdxl - pony

[hardy thorn]

create a crypto-themed logo with the name TCrypto

[brittle ocean]

This is normal for colab training

[brittle ocean]

and I'm talking as training input

[graceful lark]

smile

[quaint tiger]

fluff

[tawny peak]

bro responded to his own message lol

[quaint tiger]

hes a simp for furries

[tawny peak]

Type 97

[normal cradle]

Crypto scams monent

[shadow bane]

#✦│chat message

[surreal merlin]

Tails the Fox dressed as Jazz Fenton from Danny Phantom! Or at least, attempted to be.

Because "clever funny" joke

[leaden oyster]

Ayo? @surreal merlin level 11 !!!

[graceful lark]

Need more!!!!!!!

[arctic gyro]

Bing is down?

[shadow bane]

Nope, it’s good

[shadow bane]

It is still a bit slow though

(The site, not the generator)

[graceful lark]

A?

[forest quiver]

generate a photorealistic image of a smarmy creative director on stage with an odd audience

[tawny peak]

can’t do that here, do it here instead: https://bing.com/create

[humble otter]

bro i cannot with this fucking channel sometimes 😭

[graceful lark]

???

[tawny peak]

yeah it's trash :/

[shadow bane]

2 new animal OCs I’ve made with Bing

[shadow bane]

And a third too

[graceful lark]

@shadow bane ty ❤️

[shadow bane]

Yea

[steel moat]

https://media.discordapp.net/attachments/1229117486184530071/1244412535353118841/Default_A_mesmerizingly_celestial_cosmic_adventurer_with_eyes_3.jpg?ex=665504fc&is=6653b37c&hm=022ed1e70ead4c154a4f7dc602936686aed3838321c5a80e8c9610286029b26a&=&format=webp&width=1200&height=670

yo

[graceful lark]

wooooooooooooow 😮

[graceful lark]

😮

[quaint tiger]

<.<

hmmmmm

[graceful lark]

https://tenor.com/view/avali-gif-6297119129835963899

[leaden oyster]

Ayo? @graceful lark level 43 !!!

[graceful lark]

oh lol to be honest I haven't started making it yet xdd

[quaint tiger]

poyo made it, i just removed background

[graceful lark]

Very cute 😮

@quaint tiger xdd
Why? Why?! xdd

[quaint tiger]

LOL

[graceful lark]

What should I use for this? T-T

pony is too complicated

[magic radish]

Where do I dump the lora I trained before?

Am I doing it wrong?

[tawny peak]

What base model are you using, not lora model

[magic radish]

I trained with this setting

[tawny peak]

Ok, you trained an XL lora, and are trying to use it with an XL model? I was refering to what model you're trying to use to generate the images with

[magic radish]

the model in itself I haven't touched

which model should I download

[tawny peak]

Hooh, there's an issue with the colab XL lora trainer where it only works with the model you trained on. For example Training with XL base means you have to use XL base to generate.

[magic radish]

Okay, where do I put the xl model?

Or do you have colab for him?

[tawny peak]

here

[tawny peak]

Yup

[magic radish]

Is it true?

[tawny peak]

Gotta install dependencies first

[magic radish]

Oh, I'm sorry, I missed that.

@tawny peakIt looks like a resemblance, but it's not his face.
And the image is distorted.

He has the same hair and beard and posture.

The quality is bad and I can't put anything next to it, girl or guy.

reproduces the original image

[tawny peak]

Hold on sorry about that last text, my brother came home

[tawny peak]

Try reducing the weight, its the number next to where the lora is listed

[magic radish]

ai image

original image

[tawny peak]

also SDXL is 1024 x 1024 not 768 x 768

[magic radish]

So how do I get it next to you or in another environment?

the same person

[tawny peak]

did you try reducing the weight already?

[magic radish]

the problem was in the size.

It doesn't work because of the small size.

@tawny peak how can I increase the quality

teach me

[graceful lark]

@magic radish I think you could try using this:

[graceful lark]

• this is not bad "art"

The quality is a bit low but it's done well

[brittle ocean]

download it with base model

[neat nexus]

Close enough to Lethal Company I guess

[arctic gyro]

Anyone know How to make full anime art in bing? ?

[leaden oyster]

Ayo? @arctic gyro level 3 !!!

[neat nexus]

Take a wild guess at what these are based on

[cobalt tangle]

can someone make an fallout 4 ai art please its for my little brother he loves fallout

[magic radish]

I used it and it produces a distorted image

https://media.discordapp.net/attachments/1128747894707605624/1244782323627720724/2gK7vnaYI3LAAAAAElFTkSuQmCC.png?ex=66565d60&is=66550be0&hm=f9aa95fbd05bd7d346708cd977d9c13778a3a91147a7311762c9a29e4542845c&

For example, look at the face and the background

I am writing from mobile right now

[brittle ocean]

ye because base stable diffusion xl model is bad

[magic radish]

No, there's no problem. It's becoming very high quality.
It wasn't like this yesterday. It fixed itself.

[shadow bane]

Bro is this hungrybox or smthn

[wet hound]

wath

[shadow bane]

BRO STOP LEAVING AAAA

That’s literally the only thing they’ve said

And they joined twice

Like bro stay here

[magic radish]

https://www.google.com/search?q=tuğkan+gönültaş

Tuğkan Gönültaş

[brittle ocean]

what

[quaint tiger]

This should be easily done with bing ai

[magic radish]

I mentioned it wrong

[magic radish]

@shadow bane

[shadow bane]

Why

That ping was unnecessary

I didn’t need to see that @magic radish

[magic radish]

You said

https://tenor.com/view/ping-discord-gif-26454627

[shadow bane]

Oh

Nvm

My fault

I forgot

[graceful lark]

aaam,,,,

[leaden oyster]

Ayo? @hardy pond level 2 !!!

[shadow bane]

Fire shirt

From an Indian person

[slender elk]

just started messing with ai images again using comyui

[bronze kayak]

nice

[magic radish]

@brittle ocean

in these bouts the picture is distorted

[brittle ocean]

divide it times 2 and use hires fix with upscale

[magic radish]

should I change this

[brittle ocean]

this

[languid bobcat]

Can you also try messing with AI voice using comfyui? I added a node called comfyrvc to the node manager.

[slender elk]

Where is that?

[languid bobcat]

Install this extension: https://github.com/ltdrdata/ComfyUI-Manager

[slender elk]

Ohh that's cool imma test it after exams.. so prob next week

[tawny peak]

Trying out ArtiWaifuDiffusion

[tawny peak]

I like it

[shadow bane]

Going into surgery soon for Pectus

[tawny peak]

That's rough, good luck though.

[shadow bane]

Dawg

Stay here if you want the answer next time >:(

[graceful lark]

What do you use? (This reminds me of bing or dalee but idk)

[charred vine]

V

[shadow bane]

Bing

Also

[tawny peak]

Nice! Feeling okay?

[steel haven]

hmmmmmm.........

Is this place always quiet-ish?

Just doing a vibe check

[tawny peak]

Most people join this server primarily for voice ai as it's the main topic here. Though I guess I could say I'm active here.

[steel haven]

Ah I see; I hail from a few SD servers so I have........ about 100gb of content that isn't absolute dogshit. I like the idea of posting in new places as well. For kicks really

And because I may or may not have been muted in my usual haunt for the next 24 hours. I'll let you be the judge of the doggy doodoo comment.

[tawny peak]

Well it's fine to post images here as long as they aren't spammed or have any NSFW content in it.

[steel haven]

Sounds reasonable. There are varying degrees of NSFW though, and I don't want to toe stomp. So I'mma test the waters.

BAM cleavage!

[tawny peak]

That's ok, but the rules here state that if it's able to make people uncomfortable and they tell you to delete it, just cooperate and you won't be warned and other stchuff. :P

It's as shrimple as that

[steel haven]

UNREASONABLE!

Okay I can play ball.

Also I'm a pretty good resource for educating folks when they have questions so I don't mind being leveraged for that either.

[tawny peak]

Yeah, thanks that'll be a huge help, i'm the primary art helper so it's always nice to have more help

[steel haven]

What's your preferred UI?

[tawny peak]

I've been on and off with comfyui and sd webui forge

Though my preferred one would be comfyui because it's able to do more than most other uis

[steel haven]

Princess Xanax likes your style.

[leaden oyster]

Ayo? @steel haven level 2 !!!

[tawny peak]

I like the tech styles a lot for clothing

[steel haven]

Nice details

[tawny peak]

Are you using SD 1.5 or SDXL?

[steel haven]

I hail from the Civitai server, make models, flows.. All that stuff.

Maybe 2 loras or something

SDXL

(Not Pony)

[tawny peak]

Nice, I also make Loras, it's nice to be able to train SDXL on my own hardware

[leaden oyster]

Ayo? @tawny peak level 31 !!!

[tawny peak]

woo

[steel haven]

Yeah same here

My PC has a lot of beef and it would be weird to not use it fully

Koyha?

[tawny peak]

Nah, I stopped using kohya because it had some issues, I switched to onetrainer.

[steel haven]

I just got into it myself