#🍥|anime
1 messages · Page 47 of 1
I sent you a DM by the way
whats with all the stickers, did a sticker lora come out?
A few days ago, yeah. https://civitai.com/models/53294/twitch-emote-maker-lora
oh
Also, Vil and I got in a sticker battle earlier with our previously made ones.
/subscribe
whats the problem?
hi guys may I humbly ask what model was used for this? thanksss 😄
trying to mess around with lighting.
Stable diffusion not working
be more specific
ok show me all settings
okay
I would be really happy if I can use stable diffusion again
cuz it's really cool
need space bofre --xformers
update pip
told you in notice
as for xformers, run this in stable diffusion webui
./venv/scripts/activate
pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
pip install --upgrade pip
same
python.exe -m pip install --upgrade pip
I got to here as it said
run it from CMD
there ya go
y close that
in CMD go to yout webui folder
or opem CMD in webui folder
and type this in ```./venv/scripts/activate
pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117```
here ?
yes
all of this ?
or just from (pip)
there ya go
I'm going to wait and see if it's works
should work i think after all the updates are done
I'm an anime animtor
But 3d anime animations not cool as 2d
soo I want to do img2img for my 3d animations
so it looks more good
I did this animation
but not in stable
yes but it's not smooth
not that smooth
Why do you have every single command arg lol
I tpyed these commands as people told me
My gpu vram is 6
and my computer ram is 16
my gpu is 1660 ti nivida
I don't know if it works
Now I hope the program works
because I really like it and if it doesn't work I will have a lot of problems with ai anime art online
they need credits
Probs also have better guidance in tech support but arye and cs are asleep rn
yes they are not here
@subtle crown It's finished
but another error.
it should be named stable errors
bad news. You didn't activate the venv first so that installed to the wrong place.
run the command .\venv\scripts\activate first
just like this
your in wrong dirtectory
You'll know it worked because the next line will begin with (venv) like in my screenshot.
be in webui
oh
come on you can do it 😛
here ?
yesd here
now type pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
error ._.
sorry guys but I'm noob at these things\
so It's really nice that you help me with this
0)
pip uninstall torch torchvision torchaudio triton xformers -y
1)
delete VENV folder in WEBUI
2)
right click in webui folder and run as administrator
webui-user.bat
3)
wait for it to finish
4)
close it
5)
type in CMD in administrator
.\venv\scripts\activate
6)
pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
7) launch right click launch as administator
webui-user.bat
per line
e:
cd Ai
press tab
cd ai
cd Stable
press tab
cd stable
press tab
dir
9 - anothor error
(jk)
thanks I will do it
it gave me error when I launched webui-user.bat
is it okay?
Hello
Should I continue or there is something need to be fixed ?
hello
did you delete the VENV folder in webui
Do you help me
yes
I deleted it then I launched webui-user.bat
What is it
you said ( do you help me ) so I got it you need help with something
or you have a problem
so what is this thing or that problem?
try in CMD, python -m pip install -U pip
^do this after activating the venv again.
Wha? :o
in webui ?
Ohhh, fair then :O
or where should I type the cmd ?
I've never used that extension :o
CMD anywhere dosnt matter
Oooo understandable
Yeee, it does seeme kinda neat tho 
done
Yeee i wanna see what i can do now with this
start from step 0 #🍥|anime message
Fair, you got new power 
Gonna slap on precision full and no half and i should be set
okay I hope there is not error
After this is done ofc
i updated it, start from step 0, #🍥|anime message
that good its not globally installed
okay
should I continue ?
I know I bothered you a lot about my problem but I don't know what's happening
maybe just the staff know
or if you don't mind about helping me
your getting error becouse your installing torch 2, which is prob not compatable with your card, use pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
Did you ever check to see how much ram you actually have available? That memory error at the end of the red text means that pip was unable to cache the require files.
GPU = 1660 TI
RAM = 16
VRAM = 6
Right, but how much ram is currently available?
like this before I launch stable diffusion
okay let's see
okay another error.
ok run the CMD as administator
like this ?
yes
per line ```
e:
cd Ai
press tab
cd ai
cd Stable
press tab
cd stable
press tab
dir
then start from step 0 [#🍥|anime message](/guild/1002292111942635562/channel/1091193032273043516/)
press the tab key. don't type tab.
when I type this
it gives me this
OH LIKE THIS?
finaly
but remind me where should I typed ( pip uninstall torch torchvision torchaudio triton xformers -y )
?
goo start from step 0 #🍥|anime message
dont close the administor cmd though
yes
okay
every command in #🍥|anime message type in administator cmd
yes
.-. step 2 doen't work
why
I try to launch webui-user.bat as an administer but it doesn't open
nothing open
and I still have this open
without closing it
when I try to launch webui-user.bat as an administer, it doesn't open
check your Task Manager and see if and CMDs/.bats are not running
I'm starting to wonder if webui.bat was modified by accident.
i don't see something here
umm
I did edit something but I redit again
maybe -no half- is the problem
but I typed it because stable diffusion said I should add it
you could delete whole webui folder and just git clone a new one, make sure you save copy of webui-user.bat somewhere
It isn't even getting to a point where --medvram would make a difference. Did Chewtoy have you add the path to your python.exe in there or was that someone else?
I added python to path when I installed it
Should I install everything from zero ?
like installing stable diffusion and all staff from the zero?
and I share with you every step
I mean in your webui-user.bat file.
remove the python path see what happens, in webui-user.bat
as admin
try it
what is Chewtoy ?
it opened
now let's see
a new error ..
I think I should install everything from the zero
I will not remove this
but I will install a new one
why you trying to install torch2
run this pip install torch==1.13.1+cu117 torchvision torchaudio triton xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
Chewytoy is the guy that's been helping you
Auto1111 is torch 2.0 by default now unless you're on AMD.
ok
oh.
no it was another one
ignore that error then
i'm nivida
and install torch 1.13
IN THE CMD AMDIN ?
yes
activate first
do you have enough space on your drive?
that fine it shouldnt be a problem
i don't know what shall i do now
you could restart your PC, i dont know why your getting memoryError on the unpack
Think it could be too small or disabled page file?
i doubt he has page file disabled, as that has to be done manually, its set to automatic on every windows install
Yeah, I just can't think of another logical reason for the memory error.
start from step 5 #🍥|anime message
.-.
okay
what i think it happen is python got stock in the process when executing pip commands, restart of pc should had fixed that, hopelly, if not then fuck i dont know
ok fine skip triton pip install torch==1.13.1+cu117 torchvision torchaudio xformers --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117
ok that worked
good, should take like 5-10min
good launch webui-user
as admin?
and hope this shit works, yeah try it
doesn't open with admin
normal then
that is good, everyone get that
its almost an initiation really
huh?
wasnt easy
it was a trip is a metaphor
@subtle crown trip is not over
I get this error
and again
I have vram = 6
ram = 16
gpu = 1660 ti nivida
That's just telling you that it would need more vram than you have available. Reduce the resolution you're generating at and try again. Start at like 512x512 and see how much it takes.
it's 512x512
batch size 1?
I think it would destroy everthing
yeah that should be fine
yes it doesn't create any image
how big is the model?
thats cutting it close, try a 2gb one like https://civitai.com/models/2583 its really good
.-. I don't use the hentai model
dont matter, it can do normal art too, it just can do boobs beter
ok try https://civitai.com/models/50882/maturemalemix 2.2gb
I like this style more
it doesn't look like anime
I think I will continue with mistton
mistoon
mistoon will work its like 2gb
yes i will use it
@subtle crown still I can't create image
is there someone I canl talk with ?
Training my first lora wish me luck lads and laddies
I don't know what is lora in first
See that red icon under generate
If you click it you can see extra options add “filters”/more options to your images
I’m a newbie so I don’t have a proper answer for you
Your prompts might not be clear enough
wait this is not the problem
try restart your PC, maybe that 4GB model is still loaded into vram, when you restart pc and launch webui should load misttoon as default
okay.
Try this instead:
1boy, action pose, blonde hair, work uniform
Try just generating an image in txt2img to start with.
I think I'm going to give up. SD not working
the same problem
Might just be that you're using too much VRAM for other things. Disabling hardware acceleration in any browsers you have open and Discord can help. If you use Nvidia Broadcast, turning that off can help too (don't remember if that works on 16xx series or not). Also close any other programs that might be utilizing your GPU such as blender.
you could add this in webui-userset PYTORCH_CUDA_ALLOC_CONF="garbage_collection_threshold:0.9,max_split_size_mb:212" @fallen sundial not sure if it will help
okay
error says CPU memory not GPU memory
I run just the browser and the stable diffusion
noting else
what should I do then?
so many youtubers that aren't very into AI are gushing about Firefly. When I look at it, I'm not impressed
I assume you also have Discord open since you're sending screenshots from your pc? That and a web browser can pretty easily eat 2-3 GB of VRAM
try chrome browser and close edge
I'm using mobile
Oh ok
right now
I think it will eat more vram.
basic chrome with 0 extentions i dont think so
btw when you rent those A100 servers etc, do you get some kind of remote desktop thing or do you have to go by a cmd console?
what I don't use remote desktop or 1000 severs
servers
I live in Iraq maybe I don't have these severs
I dont know what that has to do with anything
i think I should stop using stable diffusion
it destroys my pc
I get crash a lot of time
@subtle crown i added set PYTORCH_CUDA_ALLOC_CONF="garbage_collection_threshold:0.9,max_split_size_mb:212"
Guuuuuh I hate youtube dialect....When PEOPLE speak LIKE THIS to EMPHASIZE every OTHER WORD or TWO
CAN I talk to the staff?>
guess it dosnt work
why no one suggest to me to mention them?
a new problem
hi
hi
Stability AI staff has absolutely nothing to do with the WebUI that you're using.
they can't help you mean?
You can always ask for help in tech-support but there's no guarantee you'll find anyone that knows your error or has the time over to troubleshoot it with you
you look Arabian
then what should I do?
can I really fix the problems or I should give up?
I don't know. If you give up or not depends on your personality
hey, you can try using free gpu in paperspace using lastben's repo of automatic1111
add --skip-torch-cuda-test after --xformers in webui-user
I really Need stable diffusion cuz it's cool but it's full of problems
its not guaranteed, but stable enough if you got one free gpu
okay
wait what is this exactly? is this a program I can use?
im not arabian im from UK
oh I thought that from ur name
Paperspace is a service that rent gpu to the people for a price, but they have some free gpu (limited in availability) for enthusiast to use
ok
Lastben is one among the many people who created notebook that easily run AUTOMATIC1111 in these cloud gpu
Google colab (a similar service) used to be free in the past, but it has been cracked down so I could only recommend paperspace
oh so there is a free gpu or free service it can provides?
yeah, but its very limited (not always up for a grab) since there's lot of demand
if u're lucky, u can get one for 4 hours or more in paperspace
T4 at most
just follow the instruction in the link i sent you to try getting one
Oh, Good news but...
I'm a rookie in stable diffusion, Some ace can help me with this errors? How could i find out what happened here when I switch to check point dreamshaper_5b.. ?Thanks first
Where and how can debug this error?
it seems to not be loading in dreamshaper, try the blue refresh button and select dreamshaper again
I've tried that button , But it still pop up error without any infors
I"M trying to load webuit again
Could be a bad download. You might want to try downloading it again and replace the copy in your .\models\stable-diffusion folder.
when you change models it should show something like this in the console
I've seen a few people get bad downloads from Civit lately.
@pliant epoch It seems right the as urs
what are your other settings?
Firefly is everything but anime
or western art
splash/dnd style
I did like the feature where you can modify a face with sliders though
I want the anime version of that
is there another solution? maybe the vram not my problem
When I generate it
well its good that restarting fixed one problem
It's still working. Looks like you have about a minute left until it finishes based on the screenshot.
does any images show up in this folder? stable-diffusion-webui\outputs\txt2img-images
And My Graphics card it nvidia T600, I hope this won't affect the results
Yes,I got three pigs in this folder
every single person making a preview video of firefly follows the same pattern in the video and shows the same things in the same order with very similar prompts
Three pigs at folder but it didn't show in the UI
Obvious sign of them being sponsored videos.
yeah that's weird, sorry but I don't know a solution to this
they all say the same damn thing too and a few of them shill more than the others but I barely see any of the bigger AI channels bring it up. I guess they said no
they refused to be told what to do, test, say etc
@pliant epoch Thank you anyway<:
no problem, you can try asking in #🤝|tech-support or maybe seeing if anyone has the same problem on the Github Issues tab
ok
Try three your moms
:gotem:
@lyric zealot where is smol sus
He ded... Or more likely playing a game like his status shows.
HAM
Playing games 
how dare he abandon the waifus 
I agree... Well see you guys later I'm going to play games 👋
@pastel bluff Not working either.
@pastel bluff send your mom and mom's mom, and I'll get three moms
yeah, haven't kept up with ai for like what now.. 2 months?
or less, not sure.
👹 i got life
@supple raptor @subtle crown guys thank you so much for helping me with SD
Even that I stil can't use but you guys helped me a lot
now I have this issue
It says you can disable it?
In case anyone havent seen this yet
HF: https://huggingface.co/OedoSoldier/detail-tweaker-lora This is a LoRA for enhancing/diminishing detail while keeping the overall style/characte...
I also should not abandon the waifus...
Glad we could help. Hope you get that last issue sorted.
works very well for photorealistic and anime
I did what it told me but still
to be honest I still have issues
Welcome back
careful. you're giving her an ego
Hopefully my wife won't check my discord
Nope, the gobbo girls are pure prompting on Aurora. It's an OC I'm working on. Her name is Kiki. I've been making emotes for her.

whats aurora
Luna's follow up to Kenshi. Very solid daily driver model. https://civitai.com/models/40199/aurora
Lots of gens for the emotes. The only LoRA I'm using is an emote Lora which mostly just gives thicker outlines, framing, and a slightly more chibi look. The character is 100% prompting. Here's the emote lora: https://civitai.com/models/53294/twitch-emote-maker-lora
Example Prompt: twitchemote , 1girl , green wavy hair , blue eyes , black witch hat , :D , thumbs up, okay , highly detailed , side view, fangs <...
These are Kiki without the Lora.
She made a couple of negative embeddings for it too. They're available here: https://huggingface.co/SweetLuna/Aurora/tree/main/AuroraEmbeddings
How do I use embeddings again..
If you're using Auto1111's WebUI just put them in the Embeddings folder and use the filename without the extension in your prompt. Since those ones are negative embeddings, you want to use them in your negative prompt. This is the current negative prompt that I'm using for Aurora negative_hand, (KHFB, AuroraNegative:1.1), (Worst Quality, Low Quality:1.1), border, (nsfw:1.2), skimpy, grayscale, (watermark:1.1)
It uses 1 more negative embedding which you can get here (double check that the name is the same because Civit sometimes renames files): https://civitai.com/models/56519/negativehand-negative-embedding
Well, I was napping 
I didn’t abandon my waifu!

As you may have seen from my abundance of the pink hair girl, I'm currently building a waifu
She’s locked up my basement anyway, 
FBI if you are reading this, it’s a joke
I don't believe you
I support pink hair waifu

I need to think of a name
perhaps Milly
Slave 1?
I'm not calling a waifu after bobas ship


these high res ones take so long
looks good tho
No you didn't
spoopy
aaah changed the dressings on my surgical wound to hyper-absorbent pads
It's like an Always commercial for men
Stable diffusion (with no prompting) is currently on a thigh high spree
I wonder what most people use now
original auto1111 now that he's back from MIA, or Vlad
you just press generate and see what yo uget
I mean that I didn't ask for ot to do thigh highs and it's just doing it
let the AI dream
I'm prompting other stuff but most of mine so far have had no thigh highs
Sometimes I like doing only negative prompts
I'd be pretty surprised if any of the forks have even 10% of the current userbase of auto1111
thank you for the cat gif
oh yeah GM everyone
lyrics (Drowned) by Bring Me The Horizon in neg prompt giving me and three more I dunno if I can show
@cyan burrow did you slep?
Stars don't really mean much, but Auto1111's main fork has 72k stars while vlad's fork only has 1.8k. Hell, even Invoke only has 13.9K.
vlads kinda new
compared to them at least
he's far more proactive with updating to introduce new techs
well idk how actually new it is, but it just got noticed by the "mainstream", so to everyone it is practically new
For sure, but I still highly doubt he has anywhere near 10% of the userbase that the main fork has.
oh
yeah
I heard on reddit that a ton of people like used it for a bit and then just went back lol
running Drowned by Bring Me the Horizon in Negative Prompt and "The Gunshow" by In This Moment in regular prompt
the battle of the lyrics
True, but he also breaks his UI more than Auto ever has, and has far worse documentation than Auto1111 despite most of the documentation already being done for him by Auto1111.
heh yeah
I'm going back to auto again now that he's back
what was the commandlines for getting xformers to work again?
EZ W
i think i used a bit too much de noising or something
what cfg you used?
Yeah, I must have mis clicked
I'm using my pc via Google remote desktop on my phone
So its easy to accidentally tap
oh
here is the same but less cfg
This just in: Google finally learns how to pander to the public. Can they continue on this positive trajectory or will they feel they need to cancel several projects to make up for their favorable press coverage?
Shit... It's 5am. Time for me to go to bed. Night 👋
Timezones how do they work?
is this with the detail tweaker lora?
No
do you have the link for it?
HF: https://huggingface.co/OedoSoldier/detail-tweaker-lora This is a LoRA for enhancing/diminishing detail while keeping the overall style/characte...
thx
sonar?
Uhh, ill send link
add detail on that chick in the first image makes her cheek swollen like she got bit by a dozen wasps
the photorealistic one
Also I dont know how to use Lora so I dont
ok why tf does the Detail Tweaker LoRA work sooo good man
How do I install a Lora?
download it, then move it into the models/LoRa folder
Oh I see
when you´ve done that, add lora:add_detail:1.6 to your prompt
Without the <> I assume
no
no
with
dont go overboard with it like me tho
Looks like worms
ik, thats why i said it
you can adjust the strength where the X is:lora:add_detail:X, my recommendation is 1.0 - 2.0
this is the one loading while i was asking about the lora
you can adjust the strength where the X is:lora:add_detail:X, my recommendation is 1.0 - 2.0
what model do you use
AuroraONE
Might try same but with protogen model
Or not, came out as some weird morph of 2 people
aurora is epic model
my pic from yesterday without Detail Tweaker LoRA, and today with Detail Tweaker LoRA(at 1.3):
thats impressive!
what, the 2nd or the first one?
wait isnt the left one the detailed one?
1-without lora, 2-with lora
i mean its probably my personal preference as well. but sure it might work better on other pics 🙂
now at 1.4:
then it serves its purpose xd
i guess xD
at 0.6:
question: should i put a "spoiler" on these kind of pics?
to jumpscare someone :D?
to not jupscare someone who just came into this channel
pixel art
i think its not necessary, its not that scary after all
k thx
Hi
Meow, Bond, Meow
Space surfer?
test image

gpu: crying
I think this is nsfw
do you think? maybe my aesthetic sense is a little off. i should delete?
Yes for safety
How ya'll making your images to crisp and good quality
upscaler
Also I use the tiled vae extension
yep, it was a test with a 1024x1024 using hires fix 2x
Makes it so I can do more upscale
if you're meaning the colours then I don't think you're using a vae looking at some of your posts

right I have it none, where can I get the sd uhhh something vae
skill
they are using a vae
just the baked in one
you cannot have an image without a vae
don't go into pedantics, i know all models have a vae 
download this
I'll try
best vae
hmmm where I have I seen this
then put it into this directory and refresh your ui stable-diffusion-webui\models\VAE
no need to refresh ui. just change vae in settings
care to buy a potion?

well you need to click the refresh button next to the vae setting
also in quick settings in the settings tab write sd_model_checkpoint,sd_vae so you can easily change vae
ill have all of them please
lets see if i can make (someone) do that, not her obviously
Thanks gave more color yes
and. fuck. i'm getting baked again -> my fan setup works too well keeping the system cool. except i'm in the exhaust
Spent a long time on a high res picture and it came out not as I wanted 
(not my image) 🐟 🥒
real
she comes walking to you with a suspicious cup of "herbal" teae
ill drink it if she wants me to
nice
How did you get such a good robot
What model are you using?
Mine always turn into some Gundam thing
i use Ghostmix-v1.2. great base for cyborg stuff etc.
Ah ty ty
guys stop with the jpg
....
Guys does anyone know which model can be used to generate these types of images?
nearly any of them?
bruh
you saying that like you can't just save to png
Ofc I can but I dont have to? It take a lot longer to render in png
ghibli model
what???
does it?
it does
by how much?
almost twice
ofc it depends on GPU too
👍 👍
I have 2070
i thought discord wouldnt like a 8mb png so i uploaded a 800kb jpeg
With so many models floating around and no generation info
It has become so difficult to identity which model was used...
Upgraded discord server allows higher res files 😄
took the same time
well for ya, it takes almost x2 for me 😉
nah it is fine with 8mb or even 50mb
can you try?
with one img time is the same, rendering gets longer with higher batch for me (dunno why). From 1:32 it goes up to 1:45 and gets higher with each batch
didn't prompt for a girl... did prompt for ghibli and ghibli-ish i got ^^ the cat in the water though
didnt know this. thats awesome then
night night.
Oh
Idk I only do single batches
a star
a.....jerusalem star?
I didn't know my model could do this art style too
i can't hires fix 640x640 without running out of memory
but i can gen 1024x1024 ofc
use multidiffusion
I can do the most high quality upscales on that on 8gb vram (3070 ti)
i AM using it
also, my gen speed was MUCH faster yesterday
now it's so slow
tf is happening
it doesnt increease speed it saves vram by introducing a different diffusion algorithm
so dont use it expecting to go faster
i'm not taking about multidiffusion
damnit
i'm having a terrible day,
took the vaccine yesterday and i'm feeling like shit and not thinking well
it's gonna be okaay ❤️ just rest a bit, watch a show (or sth funny) and you're gonna be alright 😉
whats the best upscaler at the moment? 4x-animesharper?
4x ultrasharp works best for me
take it easy mefistic. tomorrow is another day
a better day
eepy
covid vaccine?
yup
yeah, that one knocked me off my socks every time too
Yeah, it's always like that :/
eepy
imagine getting the real disease... be glad it's just the vaccine
stopped playing PoE for a bit, now i can put forth full GPU power for whatever i want
poe?
Fractal star ran with my vector style
its definitely hard to turn off poe
pall of euty?
oh, a good old hack and slash
trying to use regional prompt to get opponents to stare each other down but they're staring down random things
i remember that time i was on drugs with torchlight 2
aurora sucks at eepy
dfq is eepy?
sleepy without the start
lol
sleepy but cute
, so just prompt sleepy, cute
unless it has tags with eepy, it's just gonna suck at it
sle, mksks style, detailed background, solo, 1girl, white shirt, lying, sleeping, sleepy, indoors, bed
why does it suck tho? it looks pretty sleepy to me. or is eepy different from sleepy
for humans? a bit. for AI? totally
eepy is just another token altogether
because it makes sense for humans, doesn't mean it makes sense for your prompt
ok but then i also was thinking if the idea of sleepy is different of eepy, because mefistic also said it sucks at sleepy. and the picture looked very sleepy 🤔
face each other! Without photoshop!'
trying to do the whole "peaceful realm has a prophecied hero of the realm" who lives his entire life trainingand training doing nobility stuff. Meanwhile the anti-hero suffers every day, him and his friends and fellow soldiers constantly get killed on the battlefield because the hero's realm hired in several mercs etc
so they come to clash cause the hero doesn't like how the anti-hero wants to kill the bad guy
the hand is weird
faceless looking dude
gah those colors in that shape make me nauseous
Very nice
sorry about that
Now to upscale through multidiff to get detail
this is CLEAN
thx, my hentai model + detail lora at -1
detail lora is so mint
third leg 
details lora so nice
hello
weird leg
my own personal model 😄 it was an accident, but it's a pretty damn good model actually 😄
anime pixel art
that meower chilling
if the tail wasn't borked 
so cute tho! earlier i had one where a pile of tails was next to it
you can imagine why i scrapped that
oooh... i like the #2 -> "yeah, i'm a cat. i'm on fire. your point?" classic cat attitude
hey
finally getting good gens again
How comes eyes goes like this, how can I prompt it to be more smooth so when I upscale it wont look like this 
you don't want good eyes basically?
They want less rough edged eyes





