#🏙│ai-images
1 messages · Page 14 of 1
lol
@brittle ocean And I should be at the Mail where is the output folder? Or doesn't it matter?
did u even read link that i send
turn this into a gif
what
no
saved
No I don’t need it Or No I need it?!
you dont need
Hey i like this alot can you put me on to the prompt?
It looks like DAlee mini
What are your settings bro
Bro got that -5 steps
increase ur resolution
45 steps =-=
and try different weight
and download different
model
oh and
nvm
just turn off vae
For example?
Ayo? @graceful lark level 10 !!! 
😄
TURN OFF VAE
It's OFF!
Okay, but it won't change anything.
Hmm.. I like it 100%
T-T
cute, big eyes, white fur, boykisserfur
3d, realism, realistic, blender, 3d render, unreal engine, lens flare, good anatomy, looking at viewer, fluffy, clear eyes, detailed eyes, blushing, outside, red eyes
Camera Quality at banks:
The usual thing
Wanna see my settings
😄
Text To Image*
@brittle ocean Now I think that you understand why I can't read
I can Join to Voice and Show You Everything
the demons are coming
I still have it Just received it
I still can
this is checkpoint problem
try different version
??
You will Help me if this is not the problem?
SURE

How does this Colab work??!
what
New Virus:
I mean I don't mean that the Cell is Enabled
How does it work if it doesn't load?
i have no idea what u are talking about
Girlkisser?
Usually When colab is running Loading
But she's not there
Why not? 😄
nah girlkisser look like this
oh ye this is normal just click different cell sometimes cause it will disconnect u
??
[Errno 2] No such file or directory: '/content/CodeFormer'
/content
ls: cannot access 'results': No such file or directory
Download results
NameError Traceback (most recent call last)
<ipython-input-11-46e913a17b78> in <cell line: 8>()
6 get_ipython().system('ls results')
7 print('Download results')
----> 8 os.system(f'zip -r results.zip results/user_upload_{CODEFORMER_FIDELITY}/final_results')
9 try:
10 files.download("results.zip")
NameError: name 'CODEFORMER_FIDELITY' is not defined
ye u can go back to gui
😭
??
*hug*
congrats on the helper btw
Was he not?
nope
oh
Then I congratulate you too!
https://media.discordapp.net/attachments/1105964917951049859/1192824642877005853/576cd7cc0c59a8a2ab142e007b0790294b1d56cb.gif?ex=65aa7b84&is=65980684&hm=5bf 41a7956b9b92335bb051a0dd7ab3041cc95c6e428229fed8dad5cc8e71fb0&
doubt it
It would be cool if possible
A! THAT'S WHAT I WANTED
😄
I Made From Emoji Lora
We need to do the same but with Discord logo xD
Thanks again @brittle ocean ❤️
np
RealMoving 😄
Top one looks like they're actually going to have a fight in bed
Me when no boykissing
100% I Believe you
Well? Let's continue to generate BoyKissers?? :>
how do i use this ai boys?
I Use Colab

I would like to request an image of the boykisser police
1 frame before bed scene
A lil more then just Boykissing eh
the intrustive thoughts /j
Later OK?
BoYKiSsEr ❤❤❤
Can’t you just use yiffymix
Why not?
Yea but.. Its Furry My own Lora
😄
I want it to be in different styles
- Boykisser
- Furry #2
- Furry #3
And other 😅
I enjoyed doing this 😅😅
5%
10 Eph ~ 200Pictures
2%
1%
Ayo? @graceful lark level 11 !!! 
show
Yes Wait! I'm thinking how can I write so that everything is fine 😅
(I Mean prompt)
Hey pikachu
I generated some boykissers earlier
Did you see them
i guess
OO This guy ( @normal cradle ) Very HOrny @opal zephyr
Timbovill moment
M??
Yes Happy birthday
Huh
What
Yeah :))

Wat
Don't you like the way I communicate?
I'm just confused who's birthday it is 
Perhaps it’s birthday 😅
What kind of person is it like? About the Moderator or Admins or Participants
:))
idk
(My base was created in the form of png, but it’s not difficult for me to add a background)
@normal cradle What Shocking Did You See?
Me and timbo use the shocked emoji for almost anything dw
Okay.. No Problem
This is Very Good.. Even Scary
xD
@neat nexus HERE Are the Arts But You Can Photoshop It Yourself 😅
Me after I kiss a cute boy: (I have never looked like this)
Using AI to Make a Server Design Easy
Ok come on now
😭
Don't post that one
Side note boykisser kinda reminds me of changed
??! Sorry! But can you clarify what not to send?
TOOOO MUCH BOYKISSER IN THIS CHANNEL
This is what I'm Testing 😄
On second thought nevermind it's okay, dont worry about it 👍
Okay.. Sorry again
Nah it's not your fault
is there a way to generate images even if i dont have a good pc?
yea use colab
I tried using one before but it seemed broken, could you link me to a working one?
sure! wait
thank you!!!
ty again ❤️
-
Launch 1 Uploaded Library + Connected Google Drive
Download SD 1.5 Models
- Here Insert a link to the Model (Approximately: YIFFMIX or CyberRealistic I don’t remember more precisely Sorry)
Download VAEs
- It is not necessary to use
- (Sorry, I just don’t know what it is. Sorry again!)
Download LoRAs
- LoRa here
- Launch Mini GUI (I think that you can figure it out yourself / If you need Help, write!)
@mint haven Tutorial
tysm i really appreciate it (im kind of stupid)
Ayo? @mint haven level 1 !!! 
?
Is there something you don't understand?
nono i meant as in, ty for the tutorial cuz im kinda stupid so i definetly need instructions lol
Np
❤️
If you write anything, it’s not difficult for me to help
if i have any more questions later ill ask, thanks again
jungkook smiling
(Sorry it's not AI)
What is it then? Your art?
No 😅
I just found it in https://civitai.com/ But I'm not sure if it's AI
Explore thousands of high-quality Stable Diffusion models, share your AI-generated art, and engage with a vibrant community of creators
download a vae pls
tis my first tiem lel
boykisser
what tool did ya use?
Civitai is AI, it's in the name. 😂
Ayo? @languid bobcat level 12 !!! 
Yes, I know, but you can add your own picture, not just what the AI did
geezus at least downscale them a little, these blurry pixels are hurting my eyes xD @graceful lark
Pixels are outstanding because they are big and the base was made of small stickers
and other
that's why I mentioned downscaling 😛
?
like this:
TOOOOOOOO MUCH BOYKISSER and half of them look sus
average day
those u deem sus delete them plz
O.o
hi yume
:3
damn you discord
how do i visualize this
uhh, im guessing im doing something wrong here
i tried something simple and short for the first image
it didnt get into the screenshot, but she has an extra show hanging
why did you choose 1100x1100?
sadly the checkpoint author didn't specify the range of resolution in the page, anyway I can give the tip that SD 1.5 models should not be pushed past 768x768(different aspect ratios are fine), or else the image will become very volatile and you'll get more often broken bodies/limbs
you can use Hiresfix or High resolution fix to push it past that
many 1.5 models are limited to the base resolution of 1.5, which is 512x512
if you want more from the getgo, you can use SDXL models, they start at 1024x1024
since your hardware seems capable of that, I would suggest you try a SDXL model to see if that works better for you
SDXL also has more easy prompting most of the time, and your negatives don't need so many words
or use an upscaler
Gfc scale 12 that's why
Now that's just pro talk 
lmao yea
Also what's with the resolution on the right? Very confusing 😵
in normal sd 1024 x 1024
probably
is max
most of them look sus for me
What do you mean?
max resolution in normal sd is 1024 x 1024?
Higher and you'll get weird artifacts like duplicated faces.
there is no max per se, the base resolutions(so the most stable/trained resolutions) for base SD1.5 is 512x512, if the model has been trained well, 768x768 is possible
for base SDXL it is 1024x1024
then there are aspect ratios that can look like 1344x768(SDXL of course) but not every model/checkpoint will be very stable with those
many creators of checkpoints will tell you what a good resolution/aspect ratio is to be used with their creations, some checkpoints excel at portrait aspect ratios, others are more focused on the standard square one
some extra trick is the doubling of pixels, also called kohya hires fix(or deepshrink in comfyUI), with which you can start from up to double the base resolution(so 1.5 would be 1024 and SDXL would be 2048)
but I'm still not 100% sure what the resolution options on the right side of Lunas screenshot/app are
hello, can i ask something about ai art ?
when i see the development of ai art become better, i want to try to make again, but i don't know which app is better
Ayo? @empty tinsel level 1 !!! 
locally on your PC or online?
pc
you can check out stability matrix: #🔍│help-ai-art message
it offers lots of different UIs that you can install in portable mode, which means that they are self sufficient and don't need python or anything on your PC/Windows
is it free ?
of course, it's local on your PC, why would it cost anything 😉
how about civitai ?
New OC (this is my 15th one)
Pfff upscaling? Use controlnet upscale detail enhancer
well you can use that online just fine, same as bing or tensor.art
Yea!
Dowload those controlnet Tile/blur models!!
Then we get some next level "upscaling"
Then we download SD Ultimate Upscaler
And combine it with controlnet tile/blur
And add in actual new extra details instead of only enlarging and adding extra pixels >:3

Isnt that just a white tiger with a beanie in a bathrobe?
🤷♂️
XD
Ñ
controlnet sucks!!
What? Absolutely not o.o what the frick
it doesn't work how i want it to
What do you want to do with it?
how exactly do I use SDXL models? and how do i make the characters look more like the characters I want (idk if im making sense), I'e seen some really beautiful creations around so I wanna know these look so weird (especially the faces). im kinda stupid lol im sorry.
(Also im using a google colab that i was provided here, idk if thats relevant but yeah)
use a base resolution of 1024x1024
I'm not too sure what Lora is and how to install it. I've just clicked one the "play" that came on the 1. and 2. parts of the colab
..how to install dez nuts in yo mouf?* :ayo:
thanks squishy
I think Lora is like Picture Style or something like that
How would I go about installing one? and where do i find it? sorry for being annoying lol
Everything is fine! Don't worry!
- You Can Create Your Own Lora
- You can find it on the website: https://civitai.com/
Explore thousands of high-quality Stable Diffusion models, share your AI-generated art, and engage with a vibrant community of creators
so, i would just put the link here for it to install?
Ayo? @mint haven level 2 !!! 
- It’s you Forgive me If you don’t understand something Because I’m Just a Beginner myself
no its fine i appreciate the help!!
How to Use
Lora Download
- Find the Model You Need
- Looking for the Download Button
- Where it says
downloadRight Click And ClickCopy Link - Next Just Paste It into
download_lora
YiffMix This is for Furry Style I'm not sure that it will suit you
OKay so it definetly looks better now, but I tried replicating an image i saw on the lora page and it looks very different
- I don’t understand you. I’m sorry.
- Can you send your
prompt?
its on the screenshot but ill send here. I just copy pasted the promt from the first ss i sent (the first image is from the page of the Lora model I decided to download, focusing on Hange from AOT. Its what I wanted to replicate but my result was the second image)
portrait of beautiful HangeAOT, 1girl, solo, black_hair, glasses, eyepatch, brown_background, volumetric lighting, best quality, masterpiece, intricate details, tonemapping, sharp focus, hyper detailed, trending on Artstation, lora:HangeAOT:1
Can I see Which Model?
sorry, what exactly do you mean? ty for having me patience by the way
lmfao the face
she's just a little special
Link Lora
Hange (Attack on Titan) LoRA Making models can be expensive. Do you like what I do? feel free to buy me a coffee ☕ The main trigger word is HangeAO...
Have you downloaded and added to SD?
i would assume so since i was given the option on the lora drop down
Try Sampler DPM++ SDE Karras
wait nvm i read the page again, im just stupid. didnt download any models just the lora
brb
This is probably if Realistic But I'm not sure
Sorry I couldn't help you
It worked!!!
Also pls dw, you helped me a lot! I had no idea what a Lora was, tysm
Its always nice to see when people are helpful ^^
I dont always have time
So its nice to see other people taking initiative
Yes but... I don't Understand Very Much about this
- I am a Person Who needs Help, so it’s better to Help so that I don’t have to Call Another Person Again
let me introduce you to cognition A.I.
nvm thats for uhh
the wrong thing
but still anime or something
you need a pretty beefy gpu to run it anyways
and its like 2 terabytes
wait no 7.2gb
2 tb 
ah, sadly aint got one of those <//3
anyways here again to ask more question (sorry...)
-
I am currently using google colab, but I have to run/download everything everytime I open it again, is there a way for the things to be somehow saved (including the models and lora) so that I dont have to link them and install them everytime?
-
Ive seen/heard of people replicating art styles, and I'm interested in how this is accomplished!
For 2. You mean training an artstyle lora or prompting it to replicate an artstyle?
I think i mean the first one, but just in case could you explain the difference to me?
The first one refers to training a lora off a specific artstyle, the second one is just using loras/prompting to steer your way to certain artstyles
Like "by:artist" tags in the prompts
yeah i think i want the first one
Ayo? @mint haven level 3 !!! 
From what I know you would just make a lora with images of the artstyle/by the artist you want to replicate
Have you made a lora of like a character before
Its like that except when you describe the images in your "dataset" you would tag everything as opposed to tagging everything but the character
no haha, kinda new to all this
if you wanna try doing an artstyle lora on colab i can give you a guide i followed to make my own
Ayo? @autumn glacier level 59 !!! 
it has a dataset collector too
You don't need to download anything, this is a guide with online tools. Click "Show more" below. 🏭 Preamble Even if you don't know where to start o...
I would like you to try to do so yourself first. but if it really doesnt work out and things just work against ya, i can always make some LoRA models for you
In my CityPop era boyos
yo
yo
how u guys do ai art?
I do that on my own PC with my 4090, but if you don't have a powerful PC, you can also use online tools, to compensate for that 🙂
oh
Ayo? @torn wharf level 1 !!! 
I need a number to tell xD
but AMD is a hard sell for AI atm sadly
it is possible though
laggy? well the PC has to work hard to make images, so it will get slow yeah
ah okay, well again, I don't know what card you have, all I know is your manufacturer at the moment
and that is the name of model, there is no number
: (
where?
windows start bar
what's display 2?
so is mine cool? or nah
when you use AI software, make sure it uses the real GPU or GPU 1 in number terms
GPU 0 might be your CPU
it's not great, but better than CPU/CPU GPU
more power
Ayo? @torn wharf level 2 !!! 
but why i cant chnage?
you need DirectML
like it only appear cpu 🥲
wym by that?
of course
ouh
all open source code
why ppl ask here how to use okada😭
well it was kinda connected, so I allowed it for this long, any longer and I woul've sent them into the right channel 😉
Wife material
Ñ
Hello i have a question. How can i make a imagen of a carácter of my game? ( I want a pick of my carácter of baldur gate 3)
make screenshots of that character from all possible angles
make a Lora with those screenshots
Ñ
:V
What is the prompt in the Second Picture?
It was literally the same prompt as the first image.
"Bugs Bunny, crossdressing in disguise, Wearing a shoulder length black wig, pink lipstick, purple eyeshadow, pink dress, belt with gold buckle, and pink heels. full body, portrait aspect, character sheet, front view, side view, and back view.
Thank you! ❤️ 🍪
what would it cost to have a lora made ?
Free if your doing it
im not i donno how im just being curious i have my own chr i want to pose and i have not been able to figure out how x.x
For pose can’t you just use a controlnet or something?
im new i have 0 idea can you explain that to me? i have to jet to the store but ill read it when i get back just reply or ping meh i would apprecate anything u got to offer info wise =3
the idea is to take my vroid and take some chibi pictures for yt
-How much images do you have of your character?
-What is the quality of these images? (High/low)
-Does the character have a consistent look in the images?
Those are a few important questions :3
Im willing to help people make a lora when im free (currently working)
as many as you want i own it its a 3d model
4k
i suppose
Hmmm
I see
Itll probably end up having somewhat of a 3d model look in a lora
But its worth trying
my goal is just to pose her for thumbnails it just be easy if ai can do it i need a normal and achibi one
I have done that stuff before
oh i dont care about that
Generally, just make a bunch of screenshots of your character in different poses
And different angles
The best is about 30 to 40 images for the best and close results
You can however make a quicker one with like 15 to 25 images
i'll start on that tmw i guess a few examples of the type of pics u want would help i wil pm just to keep a note
this one looks so real
:3
guys what is up with civitai not showing all of the loras?, i see all these people generate with other loras, embeds, checkpoints, but when i click on one of them, i get to use em but only that, other resources that they've used cant be found.
wdym?
okay so for example i use this cyberpunk edgerunners artstyle loRa, right?, i want to pair it with the CoffeeBreak loRa but then when i go search for it by clicking add resources, it doesnt show up, why?, i see other people literally using the same resources at the same time idk why, im using civitai's web generator itself
Can you write it down and show it? 😅
(In truth, I was banned from the site, but I’ll try to Help) + This is needed in #🔍│help-ai-art Probably, but no Problems
Do not Write! Send Video
ok i will try
and uhhh u can see the past generations have been, a little rough
Hm..
its definitely seen better days, or never at all since it came out lie taht
you mean the link?
yea
oh yeah it says coffee break is not a lora, its a checkpoint so idk if that helps
--English CoffeeBreak is a checkpoint merge model. This is just a merge of the following two checkpoints. CoffeeNSFW https://civitai.com/models/437...
Cyberpunk Edgerunners (Studio Trigger) style LoRA Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or fee...
3?
uh i think the 3rd one you meant was the model?
Recommend: vae-ft-mse-840000-ema use highres fix to improve quality. 打了一个月王国之泪后重操旧业。 新版本算是对2.5d的整合,保留整体二次元画风的同时肢体上比前几个版本要好,脸型也要更多样化一点。 但光影和线条上就和2.5...
its this one in my video
You will not be able to Use:
https://civitai.com/models/94098/coffeebreak
Because it's CHECKPOINT MERGE? 😅
As I Said Above I cannot create But I will try Re-login to another Account
--English CoffeeBreak is a checkpoint merge model. This is just a merge of the following two checkpoints. CoffeeNSFW https://civitai.com/models/437...
so like what does checkpoint merge mean?
copyright
im new to this kinda stuff sorry about this lol
iz bangin doe
what
i thought you meant the music or somethin
Sorry I don't know why 😅
thats fine, ill figure something out eventually
i think
ye delete it cause it contains music with copyright
banned
Probably this should not be used on the website 😅
welp, that probably means im doomed then cuz i cant run this stuff locally loool
use colab
The model?
oh no
Yes I think why not
I'm making a Voice while I'm waiting for an answer 😅
i dunno at this point lmao
i heard colab is pretty bad, ill see if thats true
No.. I feel good with him
also idk the colab link lol
Wait
If you don’t believe my Link then you can ask the bot
-sd
does this work the same way the rvc colab does
mm 50/50
Wait, I’ll send you what I wrote for another person
!!!
https://discord.com/channels/1159260121998827560/1183928977681825883 here is tutorial
well im installing independencies rn if thats a good this
oh thanks
ye
btw is the second step essential or i can just do it when the UI's done loading (like in rvc where i can just input models)
Yes.. I'm very Stupid
You can make your own models
i have no idea what u are talking about
@sweet pewter For example:
I Want to Use this but I don't have a Button
I use it in colab
https://civitai.com/models/11888?modelVersionId=32830
The sample images are generated by my friend " 聖聖聖也 " -> his PIXIV page . I have set my models forbidden to be used for commercial purposes , so...
I Copy the Download link and paste it into Lora_download
And... I'll download it, I don't need it on my laptop
alright
Ayo? @sweet pewter level 8 !!! 
so when i conect to google drive i dont have to download the models again each time right?
M... What do you mean?
and when i download multiple other models/loras in the future, i can just pick between them?, like not to have to go back to civitai and do all over the download process
like when i close colab
and i boot it back on the next day, i will still see the same models/loras i downloaded?
You can save colab to your google drive but first write an example of your Models and lora and save and After that you will just need to click on the Buttons and that’s it
ah alright
If you Save yes
But if you just turn it off then it won't save you
also apparently it crashed while using all ram it had?
is it my ram or colabs vram
???
im redoing it rn see if it happens again lol
wym
I don't really understand you
is it my ram or colabs vram```
well, it said that it crashed while using all the ram it had, so im redoing it rn if it says that again imma screenshot
probably overloaded or somethin
cuz im using a model thats 6gb
@~@
why
idk sowwy
I haven't used it
alright
Sorry 😅
/search elraenn
Doesn't work here, though that could change at some point for stable diffusion models.
For now you can use https://civitai.com
Explore thousands of high-quality Stable Diffusion models, share your AI-generated art, and engage with a vibrant community of creators
Over 20000 total generated images!
Thanks
I like the background
What did bro generate
It was just some random person saying bad things
No image
Any bronies here?
Who wants the Mane 6/7 and Princess Luna as Sonic Characters?
I tried for Spike and Celestia but the Microsoft Designer ai didn't understand it well. The AI doesn't know Spike, and it kept generating Celestia as a Pony
Celestia too strong 
Yeah spike is a pretty generic term
Did you try something like trusty sidekick of twilight sparkles?
Hmmm
Aw hell yeah!
It worked!!
Aaaaaand, after a ton of generating and retyping the prompt differently, I finally got Celestia working
pain
O3o
11 image dataset test
You might be able to improve it further if you turn on the option: "Flip AUG / Flip Augmentation" (based on what trainer you used)
hmm i already do that
Ok then another thing you can do, is split up the existing images, by making new copies of those images, but in the copies, crop out details like the face, lower body, or emphasize details on the clothing. It's usually used more for 1 image datasets but can still help under 20 image datasets
Stable Diffusion and joe biden
Looks a bit overtrained in a few of those images, or overburnt.
Prolly I could lower a few things just didn’t have much to do in class so I decided to do that for a bit lol
Lol nice
:)
i would like to see either Goku or Gojo shred on a guitar bro that would be sick
https://www.bing.com/images/create go ahead, make it yourself
seeing the 3 years of education i took go to waste thanks to ai
(I have to work in a mcdonalds for the rest of my life)
hi
but 
i def think ai can be used as a tool for art, but thats just me
especially in marketing
Joe Biden and His Malformed Twin Brothers(tm)
Guys a question, can you use "Bing" for pc?
yes
Send me the link to download
Ayo? @latent prairie level 3 !!! 
oh you don't need to download bing to acess it
Ayo? @vivid mango level 9 !!! 
it works fine on chrome
It’s a website, so you can access it from any device that supports it
not Wacom tho
okay thats fucking cool
Wdym?
he's playing the halo 2 anniversary theme
anyone got tips on prompting?
because all i tried is just describing (a man wearing a white shirt.. etc etc) and it kinda works?
i forgot discord doesnt embed images rn 💀
"desktop"-discord, mobile works... again...
discord issue tbh
yeah the exact same situation happened not that long ago
my problem is sometimes it doesn't want to listen
you prompt a white shirt it gives a black t-shirt or somethign
Just re-convert it to png
discord likes to work
anyway, any tips?
Nah, discord's media proxy is down and images and all sent things wont load. Just wait for now.
ugh
maybe one of the fired discord workers had their revenge x)
Repost them, it’s fixed now
kk
so, sometimes my prompt works (they were meant to have gasmasks, black vests and everything else all green)
but like he was meant to have a white flecce and a helmet
and this is like the 5th image i made with the same prompt
You can try being a bit more descriptive with prompts, first try separating the base human characteristics then add a comma to separate the clothing and other details.
Sometimes AI just doesn't know what to do with large text fields and tries to fill in most things
should it be like base human, clothing 1 clothing 2 clothing 3, other details
or
base human, clothing 1, clothing 2, clothing 3, other details
Yeah that'll work, I guess i should also ask, what are you using to generate these?
the fooocus ui and the juggernautXL model
i once tried comfyui but the things it put out were some 5th dimentional shit
Personally I use Automatic1111 which fooocus was built off of, so it might have a feature that's helpful for prompting. You can try emphasizing a certain word in your prompt by doing: (your word here:1.4) to emphasize to the AI that you want that in your prompt
you don't have to use 1.4 as the weight, you can try any weight that gives the result you want
ok yeah it does have it
10 days
the number after the 1. is how much more it needs to be?
so like
(something i want a little bit:1.1)
(something i want a lot:1.5)
(something I want only a tiny bit:0.5)
Yep
Just don't make the weighting go over 2 because that's when the image can start being overburnt
discord dead?
Ayo? @earnest vector level 4 !!! 
nope.. it's working
followed the prompt perfectly, i just had to inpaint the eyes
@tawny peak Hey, i also got a problem with prompting more complex activities, like 2 people facing eachother and aiming at eachother, or one person punching another person, it just doesnt work for me
hard to do
unless you have a checkpoint that gets these scenes right, or a Lora that can force it
those things get done by others with sketches -> controlnet
or img2img etc.
do you know any checkpoints for that?

Fooocus UI doesn't natively support controlnet or img2img that I know of, it's definitely made to be simple so it's not surprising that it's missing those features.
But the checkpoint itself isn't the issue
If you want simplicity and still keep your same checkpoint, you can try out Automatic1111 as it supports LoRAs, controlnet, img2img, and other features that Fooocus is currently lacking
Initially, I thought it would be much easier to create things with AI.. but it's not.. 
well it is basically still baby shoes, and the things that people say online is hilariously misinformed, for most things anyway
right, well, it's still interesting to be living in such an era. I also want to see more of its development
yeah things will be wild from from here on out, the social media zombie era and the AI era mashing together will be gloriously ugly
If this much innovation and progress happened in just a year then I'd like to see what 2024 has for AI
Ayo? @hearty cipher level 2 !!! 


i like when she said : your ,hou and yiorkemalnge
who are u
Ayo? @plucky sparrow level 1 !!! 
Who are u
no one

now say who are u
blud what
blud ??? what is blud
fine by me
i can confirm
dalle wtf
Not yours lol
i mean....
why they want me to work 
batman
batman
@tropic pine

why cannot the ai understand basic colors ffs
what
like, i prompt a black office shirt, and it gives me a white office shirt
Ayo? @errant stump level 6 !!! 
oh ye if u are using sd try (black office shirt:1.5)
yeah i tried putting weight on it
its working for me try different resolution or something else idk
i just used gimp to color the shirt black lol
too much of a burden to make it with ai
delete this
Ayo? @brittle ocean level 40 !!! 
this is not channel for normal videos etc
where is the channel then
positive prompt :((black office shirt:1.5))
negative prompt : (white office shirt)
yeah i tried that too
send prompt
1 man (Wearing a dark grey office shirt with black stripes and rolled up small white sleeves and white collar:1.9) (wearing beige pants:1.8) (wearing a brown leather belt with a silver buckle:1.8) (Wearing a blue bandana around his neck:1.9)
try different checkpoint
1man , ((black office shirt:1.5), (black office shirt with grey stripes)), beige pants
hi
/imagine
no imagine here yet, you can use these online tools if you want: #🔍│help-ai-art message
Pov: You went to a Russian store
anyone know what to prompt to make an anime character into an irl person
Just img2img and lower weight little bit
yeah ill download a fooocus fork with img2img and try it out
Hi @brittle ocean
U can do it in automatic1111
Yo
hey btw, can i use img2img to turn images into furries?
asking for a friend
totally not making world leaders as furries
Idk
Hi guys I know I'm new hehe but still if u subscribe and like my channel I will be soo happy
Ayo? @plucky sparrow level 2 !!! 
not yet here, but you can use these links to create something online if you want: #🔍│help-ai-art message
Kalomaze is Russian 
I mean
Imagine you go to the store and are greeted by a Bear
Controlnet moment
😭
why the sugar cubes xD
Bing AI could be so much better if it just had user accessible negative prompts
how do imake art
Ayo? @untold sleet level 1 !!! 
how do i make art with ai?
Use the links I just gave you, Bing AI is the easiest to use.
if you have gpt4 its amazing probably better than others you use, or if you want something specific download a lora
Ayo? @regal ginkgo level 1 !!! 
GPT-4 and Bing AI are the same. Just a different workflow.
AI Cheated on me My gender
Thanks AI
@tawny peak
bad-artist, (worst quality, low quality, ugly:1.4), chibi, gloves, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, mutation, mutated, extra limbs, extra legs, extra arms, disfigured, deformed, cross-eye, blurry, (bad art, bad anatomy:1.4), blurred, text, watermark, multiple views, long neck, out of frame, extra fingers, mutated hands, monochrome, ((poorly drawn hands)), ((poorly drawn face)), (((mutation))), (((deformed))), ((ugly)), blurry, ((bad anatomy)), (((bad proportions))), ((extra limbs)), cloned face, glitchy, bokeh, (((long neck))), ((flat chested)), ((((visible hand)))), ((((ugly)))), (((duplicate))), ((morbid)), ((mutilated)), [out of frame], extra fingers, mutated hands, ((poorly drawn hands)), ((poorly drawn face)), (((mutation))), (((deformed))), ((ugly)), blurry, ((bad anatomy)), (((bad proportions))), ((extra limbs)), cloned face, (((disfigured))), out of frame, ugly, extra limbs, (bad anatomy), gross proportions, (malformed limbs), ((missing arms)), ((missing legs)), (((extra arms))), (((extra legs))), mutated hands, (fused fingers), (too many fingers), (((long neck))) red eyes, multiple subjects, extra heads, bad paws, bad eyes, bad head, (boring_e621:1),(EasyNegative:0.8),(deformityv6:0.8),(bad-image-v2:0.8),(worst quality, low quality:1.4),(frame, border, film grain, greyscale:1),(text, signature, watermark, username:1),(duo,female,girl,3D,realistic,CG,feral,nude,naked,animal:1),bad anatomy,bad hands,error,missing fingers,extra digit,fewer digits,extra feet,extra hands,without watermark,bad face,bad features,((bad hands, missing fingers, extra digit, fewer digits)),(extra penis and testicle:2),((extra tail,extra balls,extra hands,extra feet,extra paws))
Is this suitable for negative prompt?
O And this is not very bad, but at the same time it is bad
bit much, but sure
Thanks for the answer
- Why am I a girl
2.What's wrong with my face?!
Possibly overtrained, or not enough viable images in the dataset
Ayo? @tawny peak level 15 !!! 
The ai transgendered you 
Don't you need a Girl?
:D
Don't Think There's Nothing Scary But I Like The Way I Look But It's Not Realistic
I would like to put this on my Avatar 😅
Hmm..
@tawny peak Is it necessary that all Pictures be in 1 size?
No
OK, thank you

ok what
Normal negative prompt I think
@~@
ok ish time to try sdxl
Yo how to post image in this channel?
any good models for foocus?
focus?
ui
i forgor what type of checkpoint it uses
my issue with juggernaut xl is that it fucks up colors and things
You are @ashen mural
i just learned that image prompts are op
i couldnt get the results i needed (pose, diffrent objects) and i put a reference image in and it works

Is this your New Girl?

#🏙│ai-images A big carrot dressed as a rock star
Yeah
jesus christ that is way too big
I also have a Big one 😅
thats adorable how do you merge what did you use ? sorry im really into art XD but i suck at this stuff im really like cute things
thoughts on my neg prompt in pinned
go to the message cuz i didnt include the embeddings
a lot of brackets
I used kohya trainer's LoRA Merger
Ty I will have to look that up
@autumn moat 😦
/generate american dad full episode
Double Trolley 
(not really an ai, rendered on manga art generators and then a little bit notch of editing using microsoft paint) A rush of the blood to the head
how do you add images here
become level one then you can send images, you can't generate them directly from here though.
I can't believe it's not 64 is such a cool model.
How to create these
Confused blueberry thinking he a noise marine or smth
It took me 50k lines to generate this, can you believe it?
what in this image is ai
Ayo? @real geyser level 1 !!! 
Microsoft Designer sticker creator is available now
🤔
/create a manga
caseoh
??
Delete it and use #1159290752195633273
so true
How is this artificial intelligence used?
Ayo? @gleaming mantle level 1 !!! 
Can't generate photos in the server, though you can go to #🔍│help-ai-art and use the command /sd tools
Has anyone ever made a SD model that replicates the artstyle of Yōichi Kotabe's Mario art?
No
From what I can find, not yet. Though some one has made a Shigehisa Nakaue style lora. Which is the modern style.
¯_(ツ)_/¯
.
What?
Nothing
