#🍥|anime
1 messages · Page 278 of 1
Just how model is trained
bout the same size as animagine... probably not but if he can dodge a wrench he can dodge a ball.
Big brain stuff
try searching this channel for it, i'm sure it's been explained
thats so real
isnt that
a movie reference
Ben Stiller "Dodgeball".
YES
oh my god i havent seen that movie in forever
that ugly chick from the uhh
i forget the team name
but the chick from the 'bad guys' team
shes so ugly, she haunts my dreams rent free sometimes 🙏
mine does that often tbh
i just turn it back on
💀
if it keeps turning back on i have no issues
yeah, i don't use the vpred models, haven't been able to get them working well
I dunno, should be easy to figure it out, just throw things inside their dedicated folders and they appear on the ui
download, find your folder where forge is
lets see
im trying to remember off the top of my head
youll want to find the 'models' folder, and from there find the 'stable diffusion' folder. youll put your CHECKPOINTS there
in the models folder theres a folder called loras as well, youll put any loras there
its soemthing like that at least
just find the models folder

flips image vertically
nah itd feel weird
ive seen it like this already id know its flipped then ID flip
probably not broski if im being real
Does it have annoying girls following the MC around
no shade, just dont read much anymore
jarvis pull up how many days ive been awake
it should usually be on 7 but keep it in a range of like 4-7 i think
when i say usually i meant by default
yes
4-7 is a good range the higher values mean stricter adherence to the prompt. lower values give the model more freedom. Too high or too low and you'll see quality problems but that's what the values mean.
Then it's fine
basically youre saying 'on a scale on 1-10 i want it to follow my prompt this much'
except its 4-7
is that the animagine model?
But still, it looks blurry on her clothes and stuff, looks like something is off
out of curiosity when you run a gen what is your sys ram usage looking like?
Headset in the water 
i did that once by accident
went to the lake with earbuds
left the lake without them
we're all working to get to the point where we can tell an ai what we want to watch and have it generate if for us 😄
👍 good lucks, I don't really watch anime so
runs away
no more then delegating the task out to animators.
youve already started using ai, double down ong
AI products 

shitttt id do it LOL
money is money
$20 is $20
product is product as long as theres a customer who cares
make that bank
Ugly

yup the model is too big for you... as we feared
you can do what you want in your own home. 😆
try an sd1.5 model that's your last hope.
also maybe try lowering the resolution.
well the only thing comfy might allow you to do is segmented rendering... but just try a 1.5 model if you can't even run that at 512x512 you are totally boned.
Random magic woo
I'm leaving
I don't know forge so no idea what is happening here. how do you start forge? is there a low vram option?
Okay well let me try to put together a simple workflow and see if I can get it to run at around the target vram.
yeah I didn't even have to do any wizardry and it works with room to spare. let me find the model I used.
https://civitai.com/models/7371?modelVersionId=425083
This is the most basic of basic workflows.
just click on the image, click open in browser, and drag and drop this image onto your comfyui workspace.
It's really hard to get good quality and keep it under 4GB you'll have to just play and work really hard to find something that can work for you. It's doable but it probably won't be easy or fun... you will definitely get frustrated. I did, and I only played around with the concept for like 30 minutes or so.
Alright last attempt... you can drag this image over onto your workspace in comfyui and you should be able to run this... barely... trying to add a face detailer and a hands lora pushes it right to your limit at this resolution but this is about the best I can get on a quality level. Best of luck.
That is so basic xD
bloo sword 
Want to create Natsuki Subaru. Any good ones in civitai?
Currently trying illustrious xl.
Need accurate art style of the anime.
nice 😂
Tried something else but thanks
The fireworks I drew are a bit off, whether in moments of quiet or frenzy.What tags should I add?
Red vibe
i mean spiker, This is a terrible translation……ok,i will try it next time.
I guess it was too much still? did you start with the low vram option?
if you dont have a good gpu your options are limited
It's a struggle out here in these streets my friend.

you know.
it made me realise i made a stupid in my workflow
Can you explain to small brain non comfy users
i had my PAG values wrong
the initial PAG should be as weak as possible without creating impossible anatomy
but i had it a bit too high
yeah, workflow in png, as always ❤️
i have no secrets, and i put in quite some work on getting the PAG right -> it's specific to illustrious EPS models
but goddamn, does it deliver that way 😄
smol eliza
i'm gonna lose my status as premium upscaler when people start stealing my tech 😢
then again -> i'm pretty sure at some point, somebody will improve upon it 😄
i can see it's not working working properly, as it didn't gen a fennec. right @distant remnant ? lol
this is a comfy workflow
yeah that workflow just kept breaking, basic flux does nice enough for me lol

Tuhou 
Comfy and ugly browser
no she is cirno 
+1 
needs somewhat of critical thinking skills to use comfy without it seeming scary
Lmao
whats ur vram again
Is she not from touhouuou ?

it`s kafuu chino dressed as cirno which is from touhou 
looked the same
Baited 

KSI 
with 4gb youll be very held back
why evol
I come back later 
indeed
always evil, I'll gen some smol to un-evil you
And un-lazy
Now shower time 
least we dont gen sus images of kids like you do, weird pedo
lmao
stop projectin lil bro this aint the place for you to crash out take it to the dms 🤡
If you want, you might be able to gen on civitai?
Not sure if it’s gonna be high quality tho
not sure, a long time ago, it was already banned or something right?
hmmm, then i don't think local generation will help as well. Unless you want to generate it, then do further edits on it
if dunno how to use ? 
just 



monkey throws a dart and pops it
ong i think i finally got the style down and stable
I'm still working on this but I figured I'd post an update here. The video isn't published to my channel since it's still a work in progress but if anyone wants to play with it or use my workflow I figured I'd put together a video explaining it. It's unscripted and like 35 minutes long since I just basically hit record and started yapping.
I also realize that I never explained a whole group section. So there is a file that we save labeled json.json which just basically holds the broken down prompt information for a specific section from the last generation. <- this is useful if the ai your using is just trying out a bunch of different words and you like one of the outputs of a prompt. You can then grab that information and copy and paste it into your user defined info and stop making calls to your llm.
Using #comfyui with #griptape to create a manga. This is a work in progress but I've been discussing it on discord and I figured I'd explain how I setup some things in it.
I made several mistakes during the recording process but I just wanted to put some information out there on the project so that people could start playing with and modifying it.
between 1.5 sec and 2.5 it need a fix, for the rest, so cool
Maybe we can make our own full length anime films in 1-2 years
i really cant wait til we can feed a few character pictures in, write a short story, wait like 10 minutes and then get custom animes and stuff
my next test with kling will be seeing what it can do with video game screenshots
but yeah giving an ai a pic of goku, sonic, and sailor moon and watching a slice of life anime or something lol
reducing PAG/SEG/SWG guidance as much as possible 😄 -> results in better random details, as over-self-attention makes for dull images
now trying to get that attention good enough to fix the hands back
Looks good.
how to generate photos here?
people generate these mostly on their own computer
Or you can try online AI tools available
creative fabrica has a pretty good free online ai images, as well as the google ai test kitchen
damn, this new workflow works wonders for me
Are you using style-transfer? Or that sketchy lines are just by prompting?
just prompting
@wooden sail spykeee, yamahara, horror, girl, insect alternate species, (monsterification:1.25), eye trail, night, hair over one eye, glowing eyes, running, wind, yandere, scream, crazy eyes, claws, dutch angle, motion lines, carapace, motion blur, sideways, best quality, best detail, very awa, very aesthetic, best resolution on ntrmix
good illustrious model, took me a while to get it properly working tho
WAI v8 result is smoother. Expected
yeah, wai is very stable, but has some "freedom" issues, where'tis generaally dumber and doesn't know as many concepts as ntr does
Is there a way to inpaint an image such that the character retains the same basic hairstyle but the hair is in different positions based on the situation like wind blowing, rushing forward, etc?
lora, or ipadapter i suppose
Try red ones
red eyes?
Yes
@rigid helm
very cute gens
shame these take 3 minutes per image xD
Looks good
thats neat
and i see there's some local over-attention, still
probably high strength PAG
i'll figure it out tomorrow
idk i tried your workflow and it just kept breaking, as it tends to happen with super large and convulated workflows lol
can you tell me what's happening?
considering the workflow is so big, i cant tell what is happening to break, lol
just get failed nodes, no errors that helps
nbd, kinda expected it to break
my experience with large workflows is usually that
breakage
lol
the nodes did install?
ah. i know. it's the random artists!
you don't have those files for sure!
disable the random artists, i think that should do the trick
probably who knows, cool it works for you but as a sharing point of view this is why i dont share big workflows, its so hard to track down why it wont work for others
but i am willing to help here, i'd love for other for it to work too
thats good ^^ but im on my windows playing poe rn lol
moslty just thinking out loud 😛
mecha harambe lol
optimus monke!
bubblebee monke when
Creative
sometimes I hate discord.. I never get the fun beta stuff
just got it today, should be rolling out gradually
quack
Serene weather
no commercial models used... just a lot of work... not sure if worth it, but it isn't that bad
not sure about this one... girl looks pasted... sigh...
That would often be the case with classic anime...
but it doesn't work all that well in this case
Every artist should know:
If a character doesn't look like part of the environment it's either
- wrong perspective
- wrong lighting
- wrong details
In this case mainly perspective
A cute
So... Who of you guys run locally with Comfy? 😄
Cause i'm having some minor setting issues, and honestly wondering if there is anything i should change on this upscale.
Technically it should go into #🤝|tech-support , but this being Anime, you guys might be of better help ❤️
anyone else playing with kling or anything similar or figured out a way to get similar results on 6gb vram?
hii with 6gb vram its simply not doable
you either wait for two hours+ or out of memory error
image/test to vid is just many times more intensive
maybe LTX could work but eh its not ideal
ive gotten like some of the a111 extensions like deforum and the other one to work
but yeah the results arent great
i wouldnt actually mind waiting 10s per frame or so if the results were decent-ish
dude came to slay thots
many smol 
bloo

i want to keep genning smol 
Using kling for vid?
it looks very nice
the other day they had a deal for the standard plan for like 4.50$ so i grabbed it and it gives like 600 credits with 60 per day but i blew threw them all in like 3 days lol
All on animations?
I'm considering checking out some of these services... seems more practical than buying a $2500 RTX5090 😛
Gotta look around to see what's out there... I don't know much about any of the services. I know civit has one that seems expensive. And Kling is the Chinese one that opened up like 6 months ago.
exactly my thoughts. i tried for ages to get decent local video but im happy with images and some text for now but yeah,
my friend showed me some of the ones he made with kling and they blew me away so i had to try it and the non paid queue was 12+ hours so i figured i can drop five bucks
if u happen to find something similar or cheaper than kling def let us know
id share what my friend sent but they were all futa stuff lmao , so despite the TOS they do sometimes let u run nsfw
i havent had good results with landscape images so far though or at least it didnt produce what iw as hoping for
although he put in the windows desktop background
my thing, i was hoping the environment would come to life but maybe i prompted badly
Nice!
Spinglebob island
Will do, but not sure how serious I'll be looking. 😛
I'm a little wary of giving kling my data, though.
Any lora which shows major injuries?
bruh
How else are you going to depict a war zone?
Actually, there are so many loras trying to create beauty. But few which show ugliness
Might have to sift through a bunch of gore to make my own lora...
yeah because models trained on gore/restricted stuff isnt allowed by most platfoms
tough no idea why you would want that but okay
Well I wanna make war themed art
be sure not to share it here
I won't
Making private loras is the way ☝️
also I remember seeing cronenberg monster lora for SDXL on civitai long ago - it produced absolutely hideous images
Oh, found it. SD1.5 actually:
Titled ||Body Horror Creatures||
Do NOT look it up. For real
Not quite what I was looking for 😅 appreciate the look. Not trying to make anything intentionally hideous or gory, but more like stuff in seinen battle manga and war zones... I'll keep looking
Emergency food 
@green temple smol wizard will turn you into cookies
maybe bonk lazy cat?
So I used my agent to write a whole story and then break down all the scenes then I used that as the template to start generating images using it's suggestions and here's the rough assembly of those concepts. Obviously there's no dialogue placed into it or actual story telling here but I think it shows how you can use it to help you in the creative process.
Chapter 1: "The Canvas of Dreams"
- Introduction to our protagonist, Yumi, in her dream world, where she's a powerful sorceress named Kaito.
- A glimpse into the fantastical realm of her manga series, "Echoes of the Shattered Moon."
- Yumi wakes up, jolted by her alarm, and we're introduced to her mundane life.
Took the above and turned it into this
Chapter 1: "The Canvas of Dreams"
Scene 1: Yumi's Dream World - Kaito's Adventure
- Characters: Yumi (as Kaito), various dream characters
- Setting: A fantastical realm filled with magic, monsters, and ancient architecture
- What is happening:
- Yumi is dreaming as the powerful sorceress Kaito.
- She's in the middle of an epic battle against a monstrous creature, using her magical powers to protect her allies.
- The scene is dynamic and action-packed, showcasing Yumi's prowess as a dream-world hero.
Scene 2: "Echoes of the Shattered Moon"
- Characters: Kaito (Yumi), various manga characters
- Setting: The pages of Yumi's manga series, "Echoes of the Shattered Moon"
- What is happening:
- A montage of scenes from Yumi's manga plays out, showing Kaito's adventures and the richly detailed world she has created.
- We see glimpses of other characters, plotlines, and magical elements that make up her series.
Scene 3: Yumi's Alarm Clock
- Characters: Yumi
- Setting: Yumi's small, cluttered apartment filled with manga posters and figurines
- What is happening:
- The alarm on Yumi's phone blares, jolting her awake from her dream.
- She groggily turns off the alarm and sits up in bed, rubbing her eyes.
I didn't spend anytime cleaning it up or generating till I got the perfect shot I just kinda wanted to show my progress. There's like 70+ scenes to the story so this is only < 5% of the total but a good taste I think of where the project is at.
very nice ghost
idk if were on to the same idea but i generated a short story with llm today and was trying to think of a way to break the scenes down into prompts for images
can i run this though kling? and if yes any prompt ideas
Does someone know a good vae ?
that's basically what I'm doing in this. sounds like we're on the same page.
cyber neko is coming for you
Curious about this art style. What's this style called? 🤔
There are a lot of background removal nodes for comfy. Anyone know any particularly good ones?
I'm innocent 
Very creative. Which model?
ntrmix
i think i finally slayed the PAG dragon while keeping it alive. so now i kinda get the best of both worlds
what's your GPU specs?
3090
the final step in this workflow is kinda vram heavy
Reminding of this a bit
well, the actual prompt was flower magic \(creature\), spell, stance, quadruped ... so yeah, by all means, it should look like a pokemon
Sounds like let's keep it simple
actual prompt has the illustrious filler like best quality, and some artist tags, but those are auto-filled for me
but that's all the prompt to get that creature
and some luck
The output does looks cool.
Epic cinematic
Anyone know of a model that doesn't make the abdominal muscles so symmetrical all the time? Irl they're not perfect boxes
I Don't think this is a battle helmet of any sort >.>
*hits @green temple *

runs away
What did I do 
you didnt make enough miku's
Your Fav Color is aqua!?
REALLY!?
is it not in there ?
nope
@green temple Look Aqua
look like it so much
Not smol
winter warrior miku
Fire and Ice 
Snow will lose
No fighting
@green temple smol
Smol and Blue
Grabby 
Does anyone know if theres something like composable lora for forge?
Evil 
big 
what do you mean big 😭
not smol 
Big = not smol

yes yes, i passed my pre-k vocab test, i know what it MEANS
i just dont know why its being said 
the character
Make evil smol
nooo 

evol 
let me think of a character design
I'm making (evil:2)
proof or it didnt happen
💀
youve failed me 😔
theres no such thing as 'too evil' only such thing as 'too much of a snowflake' 🙏
thats CRAZY if you think thats too evil
maybe i should change mine a bit 💀
a lot*
yes it`s too evil for me 
im going to make the cutest character youve ever seen then traumatize them by sending them to war 
then make haizakura 
smol warrior
who
||send them to the barber and make them bald||
The :2 was just kinda broken
that's :1.1
even worse 



BAHAHAHAHAHAHAHA
make them bald THEN send them off to war
got it
its in my to-do list now 
nooooo 
I'm making a cute thing
im not :D
anyways
ive tried my hand at a smol character
back to making not smol characters
Adult character supremacy
ong
horrorurry 


i was clearly joking tho
👍
I mean its all thats being posted tho
So the ahem 9000 year olds are better then a few pictures that happen to have abs
I also post non abs
+1 
i still cant get over what chaos considers 'ugly'
Anything not "small"
the absolute disrespect to all the beautiful types of women out there-
what about the tanned tomboys, the muscle mommies, the country goths 😔
There are few distinct things that are just not cute
No
nah, gonna lose my coolif i keep going 

we're all entitled to our your wrong opinions on abs 🙏
Top 3 is frieren, ugly black hair in front girl, abs
Even when i post holo its ugly
you just named three of the most not ugly things ever
What about gura with a gigantic glizzy (hotdog) (sfw)
Delusional
glizzy is such an unneeded word LOL
who decided that should get to be a word
i need to talk to them 💀
ONG
I dont even like any vtubers and i still wouldnt call any of them ugly
Its like the obama with the goat troat
chaos HAD to have been done dirty by an anime loving woman with abs or something, i cant see any other reason to hate abs and frieren
😭
if you tattoo something on your throat you best not hope im within range of seeing it because the first thing im going to do is go "throat goat? OH SO YOUR GOOD AT SUC-" 💀
id get curb stomped but itd be worth it, i gotta let em know they look stupid 
Those are very big 🚩
Would not interact 
by god
you have the red flag sensor of a woman
shits off kilter
needs recalibrated
Lmao
this is the same vibes as "i got red flag vibes because he held the door open for me but then also held it for the woman after me"
same vibes as "hes a 10 in every way but he likes warhammer"
I mean Warhammer...
"she'd be hotter if she looked boring"
"shes a 10 but blinks too much" type beat 💀
I think this server could benefit in more active moderation because as many times you'll remind them it will fall on deaf ears 
Just grant an active person the ability to time out at least 
do mods not have perms to do that here?
We have mods?

thats kind of real
im not here enough to notice stuff like that ngl
I was being real lmao, thought its only fruit doing stuff because as something happens everyone pings her
but inactive mods and mods with little to no perms seems to be a common theme among bigger discord servers
poor fruit
doing it all herself 😔
Lately its been her, used to have more active moderation
Yeah im gonna just stop reminding and t/o or kick
I Share the Captain's Opinion On the Issue of Little Characters. There Are Those Who See them as Juste Little Character. And there Are who just them as an "Object" for realizing Their Sexual Fetishes. That's Who The Question is Specifically for the Latter - Why Do you See a sex.context in a Cute Little Character? And Lol-on, in My Opinion, IS Like a Direct Reflection of Sexacts.
Yep, feel free to send them- just not on this server
I've been really kind about it and warned plenty
I think theres a difference between, look at this anime character thats small or a vtuber etc but if all the posts are small characters its sus
The Genshin Game Contains More than 10 Little Characters. IS it Promoting Lol-Oon?
I air on the cautious side.
Being a content moderator in the past I know all too well the evils of humankind and with a tool that can activated bad actors is just something Im weary about when I see children. Im an adult, im 28, I dont come to this discord to look at images of children all day.
^
For reference, I brought this up because of someone suggesting prompting directly for "loli"
Im not here to do a morals debate, just saying it's not a place on the server for it and has been mentioned a great deal of times
No harm no foul, just generally stating 🙂
Atleast 4 this week

Could you put it in the channel description so they have no excuse when others call them out on it
Saves you the effort
Its pinned
O mb mb
I'M Just Speaking Up for The Cute Little Charters
All good, id rather not have to put it in a channel descrip to make a spectacle of it. I trust folk here 
There's a difference between posting "look at this character from game" and "prompt loli lol"
I like chibis or cutesy characters as well, but when its all that is posted with loli tags it's a bit more direct
Appreciate you Fruit ❤️
I.E. The Question Comes Down to the Fact that You canm Explicitly Indicate the Word "Loli" ... Oki-Doki
Or child or any related tag
We Promise - We Won'T Do That! 🖖
@green temple I Promized that "We" WOLDNIT WRITE SUCH BAD Words !!
damn, i rearranged my whole room and missed the most active conversation ive seen in this channel 
this one didnt come out right but its still cool
I think that just the most active people have a night on this channel.
crap! I wanted to say completely wrong! Stupid translator!
Many are now at night.

we seem to be having communication errors, i cannot figure out what youre trying to tell me 
No. No.
oh?
I missed the mark completely nevermind

lmao
Look. I have a day now. 10 a.m.
okay so its 10 am for you
Yes
its 1:40 am for me
ah yes
That's why there is little conversation
ik people arent typically awake when i am tbh
like i said, i dont sleep at night
i wasnt particularly complaining about the lack of conversation here more as lamenting the fact i missed the one conversation that happened to go on while i was awake but not around
some spanned on the ass for bad words
i saw, i was there for the beginning half lol
cant get x x pupils for the life of me, but i got ONE
ONE IMAGE
with the eyes i need

im gonna have to use inpainnntttt i fucking hate inpaint 
take away the 't' in inpaint and thats how i feel 💀
inpaint is pain
why?
dawg FR
i find it easier to just FORCE txt2img to do whatever its not wanting to do than it is for me to get inpaint to work right, and consistently
Sometimes it just doesn't give you the result you wanted lol
id actually rather spend the money to make loras to help enforce whatever concept, like a black forelock or x x eyes, it doesnt want to do
Yeah I just do this, praying that it'll get fixed with negatives lol
i DO that so i dont have to use inpaint 💀
Even after 50-60 attempts?
ONG
the issue is we dont WANT to take 50-60 attempts
im not sure about slepti
but for me
lazy 😁
I usually give up after 10 lol
i can get what i need from txt2img about one in every 10-15 gens
i know what im doing in the txt2img tab
id rather get a couple good images then hires fix them than spend the time to learn inpaint and spend all my time on hoping it gets it right in double the amount of gens
or ill just take a good image and actually draw in and edit out whatever it fucked up manually in photoshop
Yeah I usually just edit stuff on photoshop lol
I'm not strong in Photoshop
id been using photoshop before ai was a thing so im way more efficient at fixing the image in photoshop than i am at using inpaint to fix it
my specific experience in photoshop had to do with editing images so it feels more natural to edit the image in photoshop anyways
It's good when there is experience
ive been tempted to use photoshop as my ui, im pretty sure ive seen that as an option somewhere before but i dont feel like learning another ui, or adding MORE tabs and functions to photoshop 💀
i dont even use 98% of the tools and panels and windows in PS, i can get by with the tool bar on the left and the one under your layers in the bottom right
The last thing I tried to correct after generation using PS is a few fingers.
I had to do this almost on pixels
redrawing is easy if you know a few tools but ill always hate it with a passion 🙏
same with background removal
i can remove any background perfectly in under an hour but i HATE it
So I agree that the initial and good generation is 99% of success
puts me on suicide watch every time, its so BORING
I just draw to fix 
For me, this is not at all clear.
especially the giant images like ones that are 6000x9000 or super detailed ones with small particles
like, you dont udnerstand what i mean by that?
Gonna try x eyes
I mean, I don't know how to remove the entire background in PS
for the funnies
its easy
just takes time
i fucked this one up ngl
just use the polygon lasso tool
use the tool's feather functions at like .5 for softer spots, and keep it at zero for sharp cuts
eh close enough lol
i got it in this one
lets see if the next one does it too
if it does it means ive figured it out
anyways, deleted the examples of the backgroudn removal
i dont want to clutter the channel with unrelated stuff
im doing (X X pupils)
Nah they're still technically anime
Gonna make a xy plot rq
i figured i could only post ai stuff here since its an ai server
wait
waaaiiittttt
i want to do an x,y,z plot but im not noticing a 'lora' option on it
theres options for checkpoints and steps and everything else but i cant find anything that will plot loras
Banned for posting non ai 
wouldnt be the first time ive been banned but itd definitely be the first time i didnt deserve it/wasnt gunning for it 
why not normal ones
because i dont like bland and boring things
But the third option is sweeter in appearance. Though not quite what you need
Ugly 
It's the cute sparkly eyes indeed
yeah, i have a character with those pupils anyways
i want to make a few more characters with different pupil types before recycling any
im specifically making characters that might contain concepts ai hates so i can use them with newer checkpoints and loras and see how they fare with it and if i can adjust the prompt to make the concept work
And I always forget what tags should be used for this. For example, yesterday I remembered what tag makes such eyes (>:<)
I think it's just > <
In general, I need to record all the possible characters for this ...
i believe theres loras for it as well
im not gonna bother checking myself but
this guy does a lot of eye and face loras
he might have a lora for >_<
@zinc flax x-shaped pupils worked
but
im going to try a few more images
heck yeah
that was just the first try
gotta determine if its a fluke because ai does that to me
like for the longest time i could put 'highneck' and its give me a highneck shirt but then for some reason one day it decided it didnt want to work in that prompt, so i had to find something else that would 
There's no text if you don't look at it
it means i have to edit it out 
i cant not look at it
Sometimes it's model dependent so yeah
i was using the same model
i dont change my models often
like i had changed nothing
and it just stopped working
Need style lora to undumb the model 
is making a model anything like making a lora 
i feel like maybe the lora im using is the issue with the eyes

im going to have to retrain it a 9th time
or maybe just make a lora for the concept, thatd actually be cheaper because the style lora i use needed like 4k buzz to train because i used a lot of images
@zinc flax 😔
it was a FLUKE
what model are you using
It's so over
bet
What was the prompt btw
i even put x-shaped pupils in adetailer and it still said 'no'
adetailer FAILED me
What's the lora
its taking forever to download the model so i cant test it on your model just yet
2 out of 3 images came out with the eyes
no
2 out of 5
i stopped two gens midway because i noticed the eyes were wrong
3 out of 5
yeah
ill make a lora for it when i can
im just buzz broke rn
but yeah, ill need a lora for it most likely
at the very least it wont be a waste to do it, if the concept struggled with on lora im sure ill find another that will do the same meaning ill need it
for now ill just pray and gen
food on the ground
food and rabbit on table!
all food on table 🤣
Ears left ears right, the wind 
lol, the teeth
Bnnuy teeth
She is a rabbit!
yay free money
where were you when the miku androids attacked?
more miku bot
Miku is infinitely better then all these "smol" ones (we all know what you mean by those)
Thats it, once im home im gonna impaint all these ""smol"" ones bald
Miku is sacred!
did i miss smth, or is battle-bot miku needed to eradicate the smol ones?
We had fruit come in earlier about no children/lolis but creative prompting is fine while well we can see the main images in this channel
i suppose #anime is gonna #anime
It is a stereotype for sure
it was like that in 2023, and it still is now
i posted dudes too, back then -> just to equalize
Dam sadge. Here's to hoping we get more adult women enthusiasts in here instead of "smol"
.. anime explores a wide range of themes and genres, taking its viewers on a journey through fantastical worlds and compelling narratives.
"tsk, just look at this wasteland of loli"
Well, I do not like a big breast ... I do not find them aesthetically attractive.
Theres adult flat chested woman too
But only ""smols"" is icky
And I will run into the naivety and immediacy of the small characters.
the problem is that every anime girl looks young and probably is (canonically), so who gets to choose which is which?
Guess the age in this one
whos saying i'm not posting cute adults anime girls?
Did i say only you did? Im saying its the main content in the channel
i'd guess about 5 minutes or less when it was posted
Guess the age in this one
insert loli-baba argument
sure, but you can't "guess" the age of something that doesn't exist
that's not a real person
- miku is 18, as of this year
released in 2007
canonically, 16
so, at best, the creator decides, which means it could be anything
Oo the 9000 year old vampire argument
And no, the miku doesnt look like a child
Then why do you think that my invented character is younger?
miku is the perfect example of hypocrisy, 16 conocically but everyone whines about "omg no lolis"
hey, you do you 🙂
Unless all your characters are midgets? Its hard not to think otherwise
Im dying to know the prompt to get the small "women"
nearly every popular character in anime is in high school, so obviously not 18+
Do you have something against dwarfs?! They are pretty cute!
Thats teen, not loli then
i think this discussion is had every other month?
loli is from the term lolita which is a book about a 14yo
I mean the pinned message makes it pretty clear but eh
a teen
Translate loli and lolita to english using a translator
i don't have to, the origin of the word is a book title
"Lolita" Author Nabokov
but, the same people that use the term wrong also call 17yo 'children' and loli fans 'actual pedos'
Like the elon "roman salute" definitions change over time and its an widely agreed definition
they lad down due to another reduction
Anyways i gotta ride the bike home, id love to discuss further about this topic later
if you want to argue about definitions over time, might as well point out that japan, where anime comes from, uses loli/lolicon properly
Keep making them million year old midgets
Im not gonna bring up genuine loli anime examples
tbh, i'm prompting 'child' to get the cute girls, idk
Because then jts pretty clear
💀 there we go
i'm not posting lewds, so why does anyone care
if you want boobs, go to a nsfw server
guy and his frog training to take on the waifus
Lmao, anyways i gotta bike now or ill miss my train
I didn't say it had to be children or boobs, you did









