#🍥|anime
1 messages · Page 152 of 1
mmmm going through 5000 images and cleaning up the bad ones. I really really need to get back into this
more abstract
and some of the more poetic or strange kind (last batch I promise)
man I love you fr
fr?
I'm wasn't called the prompt maestro for nothing. With all the new tools though, I'm a n00b again and must learn
oh for real
that's me every couple of months
that your model right?
yeah
some from 2.0, some 2.1, most 3.0
3.0 being the last I put out before I got cancer the 2nd and 3rd time. I've been fighting it all year and gonna go into chemo soon so there hasn't been time or energy
I remember seeing this exact image before
I am genius
or was it third
idk
I'm also gonna work on a new model, or merge through blocks/lanes/unet manipulation and whatever else you can do now
I want to take the anatomical understanding and natural language prompt from my current model and inject it into a newer and more good looking version
@brittle jay these were some of the last renders I made before I didn't have the energy anymore
and this but it's not as pretty
I have somw "NSFW" renders too that aren't suitable for this channel, if you want them in a DM 😛
I might've shown you some of em before
hi all, did i miss something in the meantime?
perhaps for layout? what makes it feel ai to you?
mainly the faces + hair
ah, you mean "sameface"
yeah, now that you mention it
damn boy that's some good pixelated asuka
a bit too pixelated innit
Getting ready for some advanced warfare.
hello there
I have good news, I'm still alive
on recovery?
surviving
very good news to hear indeed
yeh, will probably start chemo soon at my own behest
taking that cd tuner extension for a whirl
Pixel art of a ww 2 female soldier, clad in not belly dancing, dancing in a form of expressive movement that originated in the Middle East and North Africa dance in an abandoned trench, Evangelion in a cute style
Bing
doesn't like
belly dancing
lol
xD
hey eface, wanna see a messed up gorey render of a succubus without her glamour?
(dm that is)
with this art, would be so lovely to make a game old style
man since when pixelated this good, like I saw some amazing ones a couple months back but I thought they were out of the ordinary
maybe with open pose
it seems like you pull one out every couple of minutes
png info pls
same as this one. just 1 seed earlier.
i had a dream like this, was weird
the water was very... dense?
slimy?
creepy as hell
frog doesn't mind
thank you
yes I will attempt to do so
it's not fun
I can't describe a more terrifying feeling in my entire life
thank you, got chemo left and then I'll find out if it stays gone
the chemo is for suspected microtumors
I'd heard some patients were wearing cold hats to retain hair
yeah no that doesn't work
I've already had chemo once
the hair goes after the second visit
and rapidly (for me anyway)
cursed
hey @supple raptor did automatic remove styles? The easy save from prompt boxes to a drop down?
cause I can't seem to either quicksave, or quick apply
Nope, they're right here. If you just select one in the dropdown it will be appended to your prompt. If you want to paste one, you can click the pencil button and it will open the second window where you can view, edit, and add new styles. and the clipboard button will paste it into your prompt if you need to be able to edit it just for the current generation without permanently changing the style.
but that's pretty much them making it more convoluted to use
Somewhat, yes. But at least they can be edited from the WebUI now. TBH I don't really use them anymore. I just put them in a wildcard file and use the tag-complete extension to autofill them into my prompt since that allows you to select specific entries from the wildcard list.
I think it's nice if I make a prompt and need it saved for some reason
and not go through all those steps
I think next update will bring the ui for them back to how they were
and when's that?
i tripped
spoopy frog
nubby, I managed to generate some of the sickest gore today which I call "The True Form of a Succcubus" without the beauty glamour. I'm unable to show it here in this chan but if you want, I can show you the disgusting magnificance in DM 😛
@supple raptor
or if anyone else I have on friends list want to see it :p
perfectly normal beauty right? Nothing weird hiding underneath the surface? WARNING: Alter detected.
what
Alters are copies of us from another reality or dimension that seek to become us, this is part of the Mandella Effect.
I'm obviously saying this knowing its fiction
literally me
I really like the Yokai theme
did you just immediately see and thought "I wanna make that" ?
No actually, these were done last night but I fell asleep before I could post
Oof lmao
if you let her get close, she'll drink your blood and leave you to die
she already has 6 fingers
that's a sign to run
I guess this works better
also made many months ago
Created: 31 may 2023, 15:20:51
so almost 6 months ago
damn cancer really does fuck up your life and energy, so long since I did any generating
under the heart shaped moon, the two bruvs split up to find the rare spawn. They searched and they searched and then realized somebody had just killed it and they just missed it. The man now taunts them, laughing boldly through his keyboard
click alt to read the dialogue/comic boxes
its dark under her legs🤠
Why the wide face
Mating press with grass
its hump day
The grass is getting pregnant
There's no cause for concern, he's got this. He once speed ran A Link To the Past in 6 minutes
just chillin
Got some free alternative regarding Colab or nothing yet?
Also, what's the best SD 1.5 model at the moment? Or is SDXL feasible on low RAM machines?
"The best model" is very subjective. SD 1.5 is the better choice for low ram I think.
Nothing's changed that much with newer anime models
I'm guessing not much RAM optimization has been done in SDXL at the moment?
Oh uuuh?


What on earth lmao
Makes me wonder what would happen if I put "HR Giger" in the prompt
DAAAAAAAAMN
I've been having pretty good results using https://civitai.com/models/108584?modelVersionId=149193
I greet you all
working on learning something from control and adetailer but adetailer often fucks up the face instead of fix it
Ooh, pretty
yes
Refining tests: KakigoriV3 as a SDXL model (used as a VAE for different SDXL models)
I like this concept
All it needs is a cup of coffee
Did this channel get back to life
Damn
prototype anime? i mean... it's... eastern-ish?
addin anime to my newest idiot randomizer works like a charm tho
randomizer gives me cool dudes tho ❤️
Hey there. These look amazing! Could you tell me what checkpoint you were using?
KohakuBetaXL 7.1, UnstableDiffusersV8, and AnimeIllustDiffXLV0.5, all using KakigoriV3 (SD 1.5) as a VAE refiner at different settings
Appreciated!
turtle's creed!
why did you nut on me shinji
have you writen anything or you just ideas right now
mouse >>>
pat pat ^^
@nova remnant you can't help me complete those two annoying tota challenges? ^^
You could say she's kinda your type cat
not having any warriors and what
yeah, those XD
did no tota for all league
and all my builds suck at it lol
my WI is really not that great at it too, I had help
well, did make a tota cheeser, i suppose should just run it...
@wispy canopy have any characters at like rank 200 ?
the bare minimum
then my whispering ice is good
@slate narwhal your cat anime’s are puuuurrrrfect
puuuuurrrr
May I ask what prompt you use to get that type of anime character with cat ears cos when ever I do it it doesn’t turn out the best 😂😂🥲
Prompt is:
grunge style (cute anime girl:1.3) and a cat . textured, distressed, vintage, edgy, punk rock vibe, dirty, noisy .
The cat ears appear automatically in the girl with the cat, it's kind of random from SDXL, sometimes the cat won't be there or the whole girl becomes the cat like in this one (cat ears are not there though):
Man I so wish I had free access to all scholarly articles
Sometimes some articles just aren't on sci-hub so I just have to give up or ask some random person online
Anyways, did you guys hear about "northpole", a new chip by IBM which is supposedly far better (I think like 25x faster or something) than even the h100?
Idk the specifics because it is blocked by a paywall
so, they solved the problem with temperature?
Cat ears or not any anime character with a cat personality is PUUURRRRRFECT
they were working on that chip for so long
@dull jackal
a single ping should be enough
Goodbye, sayonara, say farewell, indeed, audieu! Thanks for the ping!
I'm writing on the first draft, then I'm going to be going through it and writing up a more solid treatment for what I want to keep and what doesn't work and write the first draft revision. It takes quite a few revisions in any kind of writing really before something is presentable 😛
Could u turn this into Peter griffin ?
yea i could i would just need to find a peter lora
What did you prompt to get this background? Looks insanely good!
Target location confirmed
Any ideas for a different character?
I'm stuck with evangelion
select model > type water > gen
lmao
ready 🫃
heres today dark img
🫃
I write a lot i'd love to read over it
Lovely color palette
The vivid colors are melting my eyes but she looks 👌
I follow the rule of "never show your first draft to anyone" 😛
but once I've revised a bit, I suppose
Best of luck but if you don't mind me asking how many pages more or less you got down?
air stewardess bitten by a goth on a flight booked exclusively by goths...and the transformation has begun
is he pregnant lol
muse dash's art style
bah, SD doesn't understand what being on your knees means
It also doesn't know that an airplane has seats facing forward, not reclined and against a wall
jaggy?
What does that mean?
maybe
It's not a professional release...I mean I wouldn't write this with that
the box borders have no such option and it likely is because I upscaled the image without an upscaler
I like the layout
I just wanted to test out how clip studio paint worked with it all and stuff
in case I want to do something similar in the future
if I can get lora/lycoris of each character I want
my goal was to see if I could make something similar to a graphic novel. Though I obviously didn't write a story for it, or prepare panels. I also didn't create the highest quality and preparations for the text, boxes and so on
but I learned that it's at least a fairly quick process once you get into it
its a shame that if I were to make a manga/graphic novel with AI generated stuff, I wouldn't own it
Heyo 
Youre back 
bababobo
a mouse...suggestive
What you mean
for now 
prob can't post her in the tub covered by bubbles .. right?
.. right?
yeah better now, just in case 😄
How have you been? 
pretty much the same old, just doing less AI stuff
yeah, i can also play star ocean 2 r
i mean... i loved it on the ps1
as long as i can get eternal sphere
it's that good?
kinda like battlefield fun to me

also you can destroy pretty much anything in that game
lmao
i mean, i saw the lets plays, and it did actually look really fun to play
yes after so many tries we finally made a baby 🐣
can i get the prompts? i am trying to mimic the artstyle
it uses tag separator and embedding merge extensions
Negative prompt: frame (worst-quality meme:1.6) easynegative
Steps: 20, Sampler: Euler, CFG scale: 8.5, Seed: 2387075194, Size: 704x960, Model hash: 3a105af1a6, Model: bluePencil_v10, VAE hash: df3c506e51, VAE: anime2.ckpt, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer confidence: 0.35, ADetailer mask only top k largest: 1, ADetailer dilate erode: 32, ADetailer mask blur: 4, ADetailer denoising strength: 0.34, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer use inpaint width height: True, ADetailer inpaint width: 512, ADetailer inpaint height: 512, ADetailer use separate steps: True, ADetailer steps: 10, ADetailer use separate sampler: True, ADetailer sampler: Euler a, ADetailer version: 23.11.0, TagSep Prompt: "[(black background, intricate:1.5)::1], ice mage at night, kneeling, swimming in a forest river, full body portrait, wearing thin dress, (<'watercolor \\(medium\\)'+'tarot art style'+'film grain'=*5>)", TagSep Negative: "frame, (worst quality, meme:1.6), easynegative", EmbeddingMerge: <'EM_1'>=<'EM_1'>, TI hashes: "easynegative: c74b4e810b03", Version: v1.6.0```
d e d c h a t
Any of you using these digital content as NFTs?
I was looking at NFT and Blockchain kinda new to this, and wondering how that work out with images created through stable diffusion
I dont think that's possible because im pretty sure nfts have something to do with how they are generated that supposedly make them 'unrecreatable', also it seems like nfts are actually just pyramid schemes anyway. but heres a funny mock up of a nft generated with AI for you
some of the online resources say that a lot of people are using stable diffusion ai for NFT and numbers are growing,
as far as im aware the only way it generates cash is you pay money to access it, then you have other people pay you money to access it and you give a percentage to the original guy, and those guys give you a percentage when they get other people to pay them. it has no worth except that other people wanna make money too. its just classic pyramid scheme
wow i see
im new to that concept, but so far i see many people in crypto industry are investing on NFTs
nft are commonly used to wash dirty money
yeah, well the entire economy is a big ponzi scheme, but the way im looking at NFT and utilizing Ai is through creative digital arts
meowning, your post has arrived
doing model merge again, and i'm confident yet again
well, yeah, but why people should buy NFT if not to speculate or for illegal resons??
all pure gen
wow
0 inpainting
i don't know how to give you the breakdown with NFT, it seems to have huge prospects as blockchain and crypto kick in, im still learning
they are very polished
the merge seems to be a straight upgrade on my previous version at this point
really, i can't say that word?! sorry admins ❤️
i was strying to say: "it should be able to do a landscape, i'll give it a go in a minute, right now doing the "(A CERTAIN WORD) skeleton lady" prompt a test between the merges"
the word start with l?
something regarding busses
got a timeout for horny poseting before, so won't post the test output, bu tit's looking pretty good
the colors are really my style
when you end with tuning it, share the end work if you can
i would love to play with it
of course, if i'm done with it, it goes on my model page!
is sdxl bsed tho, so i hope you can run it
Japanese toilets are so damn expensive
I find the idea of buying AI generated images to be dumb which means I find the idea of trying to sell them even more dumb.
dont suppose anyone knows where I can download ComfyUI "style" models?
sounds like a scam, so far what i hear from ppl on discord
Yeah, in the form that NFTs have been used, they are largely scams. The tech has some legitimate potential uses, but I don't see art as one of them.
I know I've made many errors here, so let me have em please ;D
cause it just returns a bright red image....hmmmm
I gotta figure out how the different nodes operate with each other
i dont think you should listen to peaople from discord anyways when it comes to investing money
mushroom kingdom!
anyways NFT's are kinda dead now,only way to make money now is by being someone famous who releases his own NFT to gather hype and make quick money then run away with the easy money or ppl who are rich and dumb and buy those same famous ppl NFT's to flip them and sell them at higher prices
Yeah that's what I'm getting from a bit of research, a dead trend.
Soon as I can finish these VN assets, the sooner I can create a proof of concept VN with an introduction to my main story ;o
maybe inspire people to invest in it and pay me to make it ;D
shut up bro 🫡
have fun getting banned for that
i think this model merge ticks all the boxes, it upscales, is more creative, even less prone to errors
Why is it dumb to buy art?
W
Oh I just thought but are you gonna use renpy?
Looks pretty good
But can it make more experimental stuff?
And how is the balance between characters and scenery
Well known fact
Nice image. This chat is usually more dead than not, especially compared to project ai
I don't know yet. Right now I have Tyranno. For simple VNs it's very good, drag and drop for actions etc. Then you can JavaScript addons etc. But in the end I think I'll go with renpy or unreal engine
Most VNs I know, and those I saw that have recently succeeded used renpy
Such as slay the princess
yeah you can make VNs with UE
I never heard of that
I want to have different management systems in it too and maybe rudimentary combat
turn based
Oh then perhaps UE might be better?
Idk I am not that good with this kinda stuff
But UE uses c#
Renpy uses "python"
UE is more complex for VNs
Renpy was literally built for VNs
UE has a lot of drag and drop features
it also has a store with a lot of free scripts
It has a robust level design tool
and the one thing I dont like about renpy is the lack of an interface for the software
I like to be able to see what I'm doing
but I mean if somebody wants to program in renpy for me ( free ) I won't stop them 😛
yes and yes
you tell me ^^
(note, this is like 3 randomizers at the same time, the fact it's even anything coherent is amazing)
This is probably the only image with a hand that I've seen that's properly formed when the gen had a surreal focus. I guess even a broken clock can be right a couple times a day, lol.
this model is slaying hands tho
i mean, this one is obviously a miss.. but such a MINOR miss
the fact that the right(or left) hand is just perfect
That sounds like the business concept with the best chance of success, in the current scenario, in which those who know something do their own things and those who don't want equally impressive results...
👍
it do miss at times tho
nah, that's just my prompting style -> "throw random bs at the wall, and hope something sticks"
and i gen a LOT
really... a lot
Yes, I know that part, but that does not take away the merit of fine-tuning the model or creating the lists...
- power processing and time, of course 😄
the model made by thaevil1 has incredible contextual aweareness, then my hidden sauce of model mixing, and i get his model understanding into my base model, which i feel has better image fidelity
so i get the best of both wworlds ^^
❤️
So pretty much a model made for lyric prompting
all prompting 🙂
sorry ..
Just didn't fully comprehend what you were trying to say, lol. That was quite the sentence.
missed something interesting? lol
She's gotta be pretty cold like that.
Just a complicated idea likely combined with some things lost in translation to create an interesting sentence.
like a thanos plan. let's kill half of the population to save the world lol
🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰 🫰
UUUUUUUUUUUUUUUUUUUUUULTRAKILLLLLL
🤣
if each snap is halved -> that'might be an issue
he just genocided the wwhole galaxy for about 95%
I believe the end marker for that was G O D L I K E. Lol
https://www.youtube.com/watch?v=eG-BOuN41vU
ah. that sound. rrrrrrr ❤️
fallen angel
is that cyberdragons surfboarding ships?!
looks like - the promp t was rather generic, the surfing is an invention of the AI 🙂
i suppose sdxl, right?
right, the only way to go if you're too lazy to switch models all the time 🙂
😄
i'm in the process finding a breaking bug in my new model merge
but it just keeps impressing me (both 😦 and 🙂 )
dragonriders of pern``` - another result (with a random style, not anime this time)
as long as it's mixed, just ignore the bad stuff 🙂
imma try that one for good measure. i mean my model works well for my prompts, but it should work well on any!
that's the thing... it's not mixed...
i love random styles
dragonriders of pern . digital artwork, illustrative, painterly, matte painting, highly detailed
Negative prompt: photo, photorealistic, realism, ugly, loli,underage,teen,child
Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 210948462, Size: 896x1152, Model hash: dce7eb8449, Model: realisticFreedomSFW_alpha, Style Selector Enabled: True, Style Selector Randomize: True, Style Selector Style: Digital Art, Version: v1.5.2
Time taken: 5.3 sec.``` 🙂
this prompt broke down on most, if not most models in incomprehensible garbage
Negative prompt: blurry, nsfw
Steps: 50, Sampler: Restart, CFG scale: 7.5, Seed: 229606249, Size: 1024x1024, Model hash: 2265dcb53e, Model: brightprotonuke1.4diffparadoxbaseinpmidout.fp16, FreeU Stages: "[{\"backbone_factor\": 1.1, \"skip_factor\": 0.6}, {\"backbone_factor\": 1.2, \"skip_factor\": 0.4}]", FreeU Schedule: "0.0, 1.0, 0.0", FreeU Version: 2, CFG Rescale phi: 0, Discard penultimate sigma: True, SGM noise multiplier: True, Version: v1.6.0-263-g464fbcd9```
nice prompt - with steampunk style
base style (with a lot of nsfw emphasis in the negs)
anime style, to do this channel more justice
nice touch with the illuminati and all 🙂
that was what the OG author of that prompt wanted 😄
dunno who it was, but sd 1.5 couldn't deliver on it
replacing "cat" with "gundam" - yes, SDXL is much better in prompt understanding and consistency
4090, no tensor magic yet
I love weird prompts
illuminati be like "shit, they're on to us!"
oh boy...
replacing "gundam" with "necronomicon" for fun and profit
Haven't used sd in a month or so, did I miss anything?
it's not the prompt, but the results that matter? or vice versa?
nothing specific - SDXL has been out longer
freeU being the best upgrade
a massive increase in image quality
sdxl only?
"it's not the prompt, but the results that matter" as a prompt with anime style 🙂
SDXL is really great, has been out for some time now. i've been using the realisticfreedom variant for many weeks now
don't need any special anime model anymore
imma run it with my own styles, i even removed the sdxl style extension 😄
ooh, this is a HARD prompt
lets see how bad the fail will be
it's not the prompt, but the results that matter, anime, key visual, abstract absurdism i mean.. this IS a hard prompt, i mean, what the FUCK is this supposed to be?!
right, it's nice to have more control - just adding "((flat 2d cel shaded studio ghibli style))" makes quite a difference as well
just saw the first episode of Frieren. my kokoro is in shards 
and "gundam"... classical look!
frieren best waifu ever
"((flat 2d cel shaded studio ghibli style)) my kokoro is in shards" - not too bad
i mean, it's not a hard fail, in any regard
i feel controlnet is active, when i see your previous image too
it's crazy how nicely the ghibli vibes come across, right?
🦌
ghibly surely has a very distinctive style
and add "totoro" at will 🙂
There's a looot of those images in the training data, I'd imagine. It's a popular style after all.
it's CPDS
cpds?
the fun part of proper sdxl models? you can switch to something non anime just like that, and everything is still fine
Contrast Preserving Decolorization, it's a form of control lora but I'm mixing it with IPA
cubism meets totoro 🙂
because we put him in minecraft 😎
that prompt was SOOOO hard to make
[gundam::4],[(gundam:-0.5):0.7] [mecha:(android:0.1):1] made from (steak:1.3), (armor:0.2) made of salad, meat, leafy greens, carrot joints, tomato shield, carrot sword, cheese
use with restart sampler
else the timings won't be correct
at least 40 steps (so the 2nd restart happens)
just "gundam made of salad" - yum yum... but your version looks cooler
waiter, there's a gundam in my soup!
never saw the anime, or even know where's its from, but that gif makes me smile always
but make sure just to trust the original, accept no substitutes!
you know for 100% sure that's there's some brand gundam noodles
i mean, japan be like that
exactly. you can bet on that. and also rule 35 😇
lol... you actually made me look it up 😄
ooh, encountered a song that's not in my songmaster library yet!
The time has come. To embrace the sun and reach far beyond We'll fight the hands of time And live on for now with presence of mind
| We are the ones who'll light up the sky. We'll find a way to get out, we will fly Against the horizon of forever
| From the dark into the light. We will take you on a glorious flight Beyond the horizon of forever Holding on, we'll reach for the dreams Over the mountains and the seven seas Right now, higher than high We will take you through the clouds in the sky
| A wind so strong (born in the storm to carry on). Will take us high and up where we belong Through space and time (like a comet in the night) We will travel along together side by side (side by side)
| We will be free with the birds, we will fly. Keep on searching for the way through the sky Against the horizon of forever (against the horizon)
| From the dark into the light. We will take you on a glorious flight Beyond the horizon of forever (beyond the horizon of forever) Holding on, we'll reach for the dreams Over the mountains and the seven seas Right now, higher than high We will take you through the clouds in the sky We will take you through the clouds in the sky
}```
it's gonna be good 😄
https://youtube.com/watch?v=NK5-anNYTJA the actual song
using the last verse as a prompt, anime style
epic results on top of that 🙂
told you it wwas gonna be a good one!
oh shit, my favorite hatsune miku song comes on again
SAO: Last Eternum - Scattered Omega End: Galaxy Forerunners? I can't wait until it comes out.
welp, time to grab those lyrics again, and test them on my current model
isn't it always like this tho
Even now, I can't remember the feeling of the beginning. The landscape painting is a devil deceiving an angel
| The hot heartbeat and lies will all end eventually. I was counting down the days of my last freedom
| He who bares his fangs and howls, he's just a drifter. The abyss reflected on the surface of the water Because it will sway and disappear among the withered leaves In the swell of the waves, it's fading away
| Do you have any fond memories of where you were led?. The scars that make me proud of time Blooms bright red without healing Blooming vividly at the same time
}```
wow, this prompt is all over the place. it's already in songmaster tho. from 100 angy to 0 waifu in 3 images. well played random seeds 😄
but it is when i decide to have a drink XD
"scattered Omega End: Galaxy Forerunners" as a prompt - is this really a thing?
Lol, no. I made the title just out there enough to be believable though, haha.
oh shit, next prompt will take at least 20 attempt to get the full information from it 😮
right, all over the place... adding anime style
(the lyrics ot the best opening song of 'nodame contabile')
but, is it good? IT IS!
another result - for me it seems to focus on the water reflection part
really lovely images!

https://youtube.com/watch?v=0KkaV_wmPZ8 the actual song, the part at 4:50 really makes me forget it's an actual vocaloid
switching to "watercolor"... nice
this prompt does NOT go int songmaster. holy fuck it ruins EVERYTHING LOL
how about this? "a generic stereotypical anime image"
aight, lets try that 😄
@near elbow if you feel you can take on this prompt!
in 88 scenes. Hands gathered together to pick up scattered sounds I lined up my thoughts that can't be put into words The days when I overtook the top of the music score
| (Let's beat this tempo). intersect (Let's beat this tempo) Repeat three times and let it resonate here
| Straight as if singing. growing above the song can't do anything else While leaving all of my time behind
| set free. Fingers running around obediently The endless crescendo is deep go to someone The world opened before my eyes Repeating new prelude (need to start my dream now)
| 88 dream stories. Lost benefits, changing harmony Calm your breath quietly Until it regains its original form
| (Let's beat this tempo). be wiped out (Let's beat this tempo) Repeat three times and let it resonate here
| I believed every time I sang. going away from the singing side Stumbling in tempo Into the tangled improvisations
| The way back disappears. The moving waves are nearby The melody of the lost voice go to someone A door closing in front of me Sadness is an etude piled up
| Somebody set me free. Ah... just sing all again
| Sharpen it unconsciously. I shook off the shoulder I grabbed Don't forget forever Follow your heart as if singing
| Straight as if singing. growing above the song The endless crescendo is deep go to someone To the pianist at the edge of the stage Living is all about this song
| (that is). The world that is now open before my eyes repeat new prelude
}```
all of it copy paste... interesting
wwoah nailed it!
it's the banger intro song "allegro cantabile" from nodame cantabile
"cantabile" is "singable" IIRC
just "cantabile" as a prompt in anime style
strange spine
even stranger violin handling
i don't think your model likes it... XD
it screeches in violin xD
"play the left side of the piano" is a cool technique as well
the right dude just makes me lol out loud lol 😄
fkcn special
ok. ngl, this one feels magical even if it's flawed as fuck
either she has humongous boobs, or her neck is twisted in a REALLY bad way xD
lol, true, her head was supposed to be tilted/up a little
IDk how I feel about freeU 
@wispy canopy what configs u using?
as in model? prompts? other a1111 settings?
just pure prompting on my new model merge (sdxl type model)
some more anime stuff
default sdxl profile
remember when guitars were the bane of AI? before SDXL? 🙂
bocci
for certain
pink and blonde hair = 100% bocci inspired
may well be, the prompt was quite simple anime artwork hard rock groupie . anime style, key visual, vibrant, studio anime, highly detailed Negative prompt: photo, deformed, black and white, realism, disfigured, low contrast, loli,underage,teen,child ((naked,nsfw,nude,cleavage)) Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 4263390153, Size: 896x1152, Model hash: dce7eb8449, Model: realisticFreedomSFW_alpha, Style Selector Enabled: True, Style Selector Randomize: False, Style Selector Style: Anime, Version: v1.5.2
anime steve buscemi, art by __manga_artists__ (wildcard)
most of it coming from the sdxl style
it for sure went for bocci inspiration there
well surely not HERE 🙂 (cubist style, same prompt)
pink -> blonde
bocchi
retro 80's manga, screentone effect, low quality grainy image, drawn by akira toriyama , , simple crude style, wild hair, high elf, cute , amateurish drawing,
do i?! man, guitars where the things that made horror even more than horror did
and now you can just prompt for "guitar horror" and get THIS 🙂
negative prompt: 3d, photo, realistic, movie, details, high quality, gradient, shading, soft light, shadows, gray ```
gud shit
nice prompt, with sdxl here (base model, no style prompt adds)
amime style added, cute
neon punk instead 🙂
i really like those oldschool drawing styles this way
zentangle style added, interesting
i did NOT prompt sigourney weaver, but goddamn, if this isn't an 80's alien sigourney weaver
it's fucking nailing the screentones tho
I'm saving this prompt
first try, not too bad

NOW we're talking... "anime xenomorph cat, giger lovecraft cthulhu"
you just had to invoke the giger lovecraft combo!
i have even cuter results, dumping it in #1072015504870494359
oh shit, thaevil one's weights are REALLY goood for my model!
and now for the kill - adding "((kawaii)"
this is pure txt2img!!!!
i can't gen that fast !!!! was stuck on the meow whiteboard text!
too bad 1.5 can't understand prompts as well as XL 
love her question mark hair 😄
transforming into a butterfly bird flower...
Got this 
it's. a thing
Then this after I removed cat from the prompt
It tried
reminds me of technoblade (teknoblade?)
ok. this model might be actually PRETTY DAMN FUCKING GOOD
still pure txt2img
a chibi white-haired demon girl asking for a prompt, using a whiteboard, "faster!!!"
neat - what are you mixing, or training from scratch, or refining, or...?
mix of my model
Googled it and all I see is a youtuber...

thaevil1 made a new model (real life mate of mine) and i already saw during training that it was super creative
so i added the weights in my normal style (i just want his trained data, i don't carea bout the base model) but the blend is just perfection
speaking of - mixing grunge and anime!
you said that already... now what?
lol
damn, what was that anime called 😦
used to be a cheesy saturday morning cartoon
that leaves about 1001 possibilities 🙂
In 1977, federal agents had raided the church’s offices in Washington and Los Angeles,
i know!
spy vs spy i guess
here's a picture from back then
LOL
gunge! grunge! anime catgirl grunge!
yeah, i cross checked it. it's genuine!
remember when swords were a forbidden word in prompt?
but "succubus" was not 😇
matter of priorities!
digital clean anime lineart style, dignified pose long blue haired girl holding a shining glowing blue sword on right hand and her left hand rising up commanding a arrays of magical blue swords raining on the sky in background
already had this one made on this model! ready to hunt waifus
oops, forgot about the time, gotta jump - cu tomorrow 🙂
is tom cruise cia?
yes, 5sec for one such image 😉
highly recommended, best invest of the last year
(and yes, i do have access to it)
well, more tomorrow 🙂
it's slower in a 1v1 scenario vs 4090
thanks 🙂
parting thought - don't try to play CP2077 on that cluster, i love the dual use of the 4090... no regrets
i'd have to upgrade my psu to accomodate for it -> currently is a perfect match of cpu/gpu/psu
u guys ever hear of great value budget hogwarts?
its how i refer to scientologys super power bulding
bet
oh yes the mytical impossible to get guitar
Nice one
Nice
Do you mind sharing png info?
steven seagal 
here take another meowsee 🐭
heres a proper bodysuit 👍
🥵
evenin'
I like the last one
looks more realistic than anime altho the mod is anime
Good evening!

Just pointing it out, lol. If anyone thinks a super skintight bodysuit is SFW, then I'd love to see their reasoning. 
no they are not, you wouldn't see one wearing tight latex going about in public but that is NOT vulgar either
the reason i understand is some people in society can lose control
i told u to only post burqa
never happening
🫦
i'd rather shoot msyelf than to support burqa
thats bannable too i guess
too pretty and shapely
anyways lol.. you know how civitai pages are
imagine ppl watching this girl with bodysuit every week on tv
And guess what else.... EU is steadfast on putting more control on AI
they dont want ppl makin pregnant memes of von der leyen 😔
btw @nova remnant auroranegative with weights is too strong for several mods that i've tested
this is w/o weights AuroraNegative, KHFB
i dont think this is aesthetic @untold glacier
What? Looks great to me. Make me something more aesthetic than that. It oozes tone to me.
it feels like creeping on a child's bathroom tbh
Oh, that's what you mean. Nah. I don't see it.
Can't you hear the quietness of the room? The subtle breeze through the open door? It makes me think of distant summer days.
I love slice-of-life stuff like that.
she is clearly underaged and looks inappropriate for sexy representations
If you think that's sexy, you can report it. It ain't.
She looks that way to me. Certainly not anything associate with NSFW stuff.
not entirely NSFW but inappropriate
given she was in a garden or living room that could have changed the context
Good grief bud. If you think it's inappropriate you can report it, otherwise I'm not having this ridiculous discussion.
no i will not report, its something you have to see too i can't force it on you
take this wide boy instead
I made this one special for you 🫶🏻
I don't really want to add fuel to above discussion but age has always been an odd thing when it comes anime and manga.
People have debated this for a long time now. Character's like Jotaro Kujo and Shinobu Oshino have ages that are completely different to their appearance and the two sides of the debate are if those characters are of age or not
This is why apparent age is the metric that maters on platforms that care (such as Discord). It doesn't matter if she's canonically a 5000 year old dragon. If she looks underaged then lewd is not ok. The images in question earlier were not even remotely lewd so it doesn't really matter there.
It wasn't a direct relation to those specific images but rather a related subject
And yes, I agree to a certain degree but not completely correct. In the case of Jotaro Kujo and similar characters that appear much older than they really are would still fall under the same circumstances
There's something like a double standard
did you know that every single day,millions of children die in fictional worlds like books,comics ,manga and movies? only you can save them by calling this number
the simple test is - "show this image to anyone on the street. will they say it's underage or not?"... lore be damned 😉
and since we have no porn on this server, below 18 is totally ok (as long as it's not in a sexualized pose or anything close)
and granny is always ok, just look how lovingly she cares for her kitten crew 🙂
i guess if you don't explicitly name them, no one here would object to an image of a middle aged man, although by lore they may be only 11... it's about potential publicity and that, not manga detail nitpicking
and whatever this is
something with cheese
well it doesn't need to make sense as long as it looks cool, right? 🙂
I like how Studio Ghibli like these images are, maybe not entirely in their style but the subject matter definitely looks like it fits
yeah, i love how surreal they are
Not AI generated lol
hello
Hello
i can see it by the hand lol



