#💬｜general-chat

1 messages · Page 34 of 1

rapid plover
#

Either way hoping to absorb as much as I can and rly make sense of this technology in the coming months.

carmine otter
#

(╯°□°)╯︵ ┻━┻

#

Não consigo mexer

#

Não funciona

wide juniper
#

The symbol '|' seems to merge my words. Isn't it supposed to produce separate results one by one?

narrow sluice
#

Jae, Is Stable Foundation a new thing?
Feels like when I read the announcement, it said you guys are setting up a Stable Community, but it wasn't terribly clear from the wording that that's why it's called Stable Foundation, thought it was part of the sentence since the Community part was bolded.

#

And I didn't expect you guys to add your other projects in here, which really threw me through a loop.

#

Also, did you change your logo too?
I'm seriously overwhelmed by how much you guys changed, ngl.

wild steppe
narrow sluice
#

I almost thought I was in a new server, when I clicked on it.

#

Since all the chats got streamlined or moved down

wild steppe
#

I can see how it seems confusing since it happened pretty fast! We're around monitoring chats if people have questions so hopefully it'll become clearer with some time! If you have any feedback, feel free to drop it in #1002293361526460608 !

narrow sluice
#

Thanks, Jae

shy thicket
#

Yeah I did wonder what had happened lol.
Anyway I have a question.
Is there any news on the next SD update? Will we get SD 2.2 or will we move straight to SD 3.0 as let's be honest the uptake for 2.1 has not been great. Virtually every model I see created is for 1.5.

runic mortar
#

our society isn't stable at all

spring surge
#

is it possible to colorise a black and white photo with SD?

odd stag
#

how would i go about training a new concept into stable diffusion and further train it based on that concept

trail wing
#

Seems like someone announced something and someone twisted their knickers so hard, trying to show there's other stuff and just not image generators. In process they just killed posted generation channels. But hey, who cares.

warm junco
trail wing
#

nah, the channels got archived and the new ones are zombified now

#

plus you need this weird interaction to pin and see as full channel

#

so nice shot to the big toe, anime probably will survive because of thirsty dudes but that's it

shrewd grove
#

I think Super Stable Diffusion 2.0 was just Aitrepreneur's somewhat clickbaity name for Automatic1111, as it was such a game-changing new UI. So you're all good! If you want to branch out from 1.4/1.5 you can also check out fine-tuned models from the community (e.g. on the Civitai website) which can make it easier to get good results in certain styles!

shy thicket
#

Do you need more than 8GB to make embeddings in 2.1?
I can make them fine in 1.5 at the moment with a 2070 8Gb.

fervent thunder
#

Is Dreambot free?

#

Can i implement it in my server

vapid bone
#

Hey peeps, trying to wrap my head around all of the possibilities before I try getting setup and running. Is there anyone here who is prepared to give me a run-down on Stable Diffusion environment? I've seen various references to Automatic1111, Deforum and a few others thrown about. What are they?

Happy also to just be referred to some good reference material to read up on this.

I'm pretty sure I've got the hardware needed to run this.

warm junco
vapid bone
warm junco
#

Correct, Automatic1111 has Community Extension support so Deforum can be installed as extension

#

Can link you some sources there

vapid bone
#

Love it, thanks some much for your advice. Installation tips is exactly what I need.

strange lily
cloud granite
#

I'm leaving this server. I honestly gotta say I don't really like these changes... Goodbye

charred notch
#

is there a generated images database i can search through with promts?

simple plaza
#

Is this the official stable diffusion server ?

near falcon
#

how do u copy and paste prompts?? from image

real zodiac
#

the new icon looks kind of... soulless umm

restive prawn
wide juniper
#

How do I delete a style?

fervent thunder
#

Is the bot still here I joined to get the bot in my server

#

Lppppfpq

wispy jasper
naive schooner
#

hi everyone, happy to join and share experience with you about Stability AI

molten bough
#

SAme same, mostly here to ask questions haha

timber drift
#

Is there any channel here to post images that aren't necessarily pretty, but make one feel like "yo wtf is going on in that picture"? I mean, confusing diffusions but not really by mistake, like, weird scenes that dont make much sense without context.

wild steppe
cedar flicker
#

Advanced Topic: Negative Prompting [New in DreamStudio!] I am testing dream studio
but I can't see how the negative works, who explains a little better help

nocturne meteor
#

can anyone point me to the easiest way to create your own face-model? I tried dreambooth but it was very error prone and the output was semi-okay

steep wing
#

My pfp is a LoRA trained on my face, in 2.1 but 1.5 should work if you still use that

nocturne meteor
steep wing
nocturne meteor
astral goblet
#

I just want to say that Foundation is a great scifi series by Isaac Asimov, and that the server name change towards that is some super scifi bad assery.

Can't wait for season 2 of the long overdue series

jovial tree
#

Elo

nocturne sail
#

ey

jovial tree
#

Who are you?

#

Are you who?

#

Are am I?

nocturne sail
#

me is you

jovial tree
#

I think therefore I am

#

No
I am yu

nocturne sail
jovial tree
#

He is me

nocturne sail
#

lmao

glacial glade
#

heyaaa

swift snow
#

Where are the genre channels???

wintry stream
swift snow
#

Ahh thank you! Really confusing at times

kind sinew
#

Some artists may view AI as an instrument wielded by large corporations, which has led to resistance against it. However, the recent opposition from companies like Google towards AI models like ChatGPT raises questions about the stance of AI development organizations like OpenAI.

Is it possible that these entities are actually working towards combating big corporations, providing support in the ongoing struggle against them?

It seems like by rejecting AI, the artists in opposition may be inadvertently limiting their own resources in this fight.

quick bobcat
#

Longshot question: anyone know how A1111 curates his online extensions registry? My extension became unexpectedly popular and it seems like the right way to go.

tardy escarp
#

I think, therefore I am not.

fervent thunder
#

@delicate oxide

delicate oxide
#

whats up

narrow sluice
astral goblet
#

asimov tends to be brilliant, yet dry

#

most science fiction from those past decades in fact. it wasn't really until crichton that scifi started getting the hollywood treatment

proper karma
#

ayo wtf did i miss

thorn sigil
#

Does anyone have experience with the Animator v6 extension on Automatic1111?

forest apex
#

yea my animations often go weird though

thorn sigil
forest apex
#

yea do you get an error or it just ignores it?

thorn sigil
#

It just ignores it

#

I've tried with commas, without, trying them on the same line, separating lines but it just seems to focus on positive prompts

#

when I do img2img with the same prompt it produces the result im looking for

forest apex
#

so you're trying to do it from txt2img?

gleaming cosmos
#

woo-hoo finally got my tesla m40 24gb working! Loving the VRAM, it is faster than I expected. An image on average every 7 seconds(default settings) with batch size of 8

#

now I can finally do training

#

not bad for $200

thorn sigil
forest apex
#

you can also change the denoise, etc on another line

thorn sigil
#

even if I change the denoising strength it still completely ignores what I wrote in the keyframes

#

I've ran it at 0.3, 0.5, 0.7 with no success

forest apex
#

how many frames between keyframes are you waiting? and how many fps and what length is the total video

thorn sigil
#

So to test the keyframing quickly I've been doing 15 frames 1s 5 frames between, yesterday I did 300 @ 10s

#

the correct format would be frame# | prompt | positive | negative |,
frame#2 " "?

#

not frame#, frame#2 | prompt1, prompt 2 | neg 1, neg2 |

forest apex
#

yea on each command on a new line, and you dont need to add the last pipe if you don't have a negative prompt

#

0 | prompt | a garden | 5 | prompt | roses and daises |

thorn sigil
#

no comma right?

forest apex
#

no, and prompt is the command, it depends on what version of animator you're using? the latest extension version uses that format

thorn sigil
#

yep I'm on v6

thorn sigil
feral mist
thorn sigil
queen hatch
#

probably a dumb question but is there an Image to image AI generator that I can use here for free?

thorn sigil
#

I think playground ai does img 2 img

#

free 1000 imgs per day

queen hatch
#

thanks!

steep timber
#

guys

#

is there a model/embedding that is specialized in generating clothes designs?

dawn basalt
#

Whats the best group of ai tools to use for productivity

mystic remnant
#

i'm trying to launch SD on ubuntu with the API enabled. i added --api to the command line args file but no luck. anyone know what the deal is?

quartz cape
#

no more favorite images?

vast pollen
#

I've been getting weird purple spots randomly on results with anything 4.5

astral goblet
#

the authors hate that v3 is so popular for whatever reason. they're out to displace it by adding porn and other hard anime merges into v3, calling it v4, but what they're really accomplishing is blown out vectors

orchid prism
#

anyone know of a tutorial / steps to upload a photo and then have stable diffusion generate something in that photo from a prompt?

orchid prism
static agate
#

Hi guys!

#

toucan bird is my favourite

gleaming cosmos
#

Wow I got my Tesla M40 working and it is great. Was not sure what to expect for so cheap but it can do all of the training types. And I expected it to render a lot slower but because it can do 8x batches it is actually decently quick. About 8 seconds per image on average

gleaming cosmos
mystic remnant
gleaming cosmos
#

ya

mystic remnant
#

wheres it located?

mystic remnant
#

it uses .sh

gleaming cosmos
#

ya you edit the script that ends in -user.sh (forgot the ful lname) but you run it with the regular script

#

it calls it internally

#

it is the same regardless of OS jsut edit the sh instead of the bat

jade mesa
#

Why did y'all take the bot away

delicate fossil
#

what channel do i go to, to enter prompts?

mystic remnant
warped vault
#

hey does anyone using pix2pix know why it isnt showing up on my anything UI? Are there some i need to add in the quicksettings or extra networks area?

mystic remnant
gleaming cosmos
#

I am not in front of my install right now. There should be one that is called "command line args" or something like that

#

that is where you put the command line arg you want to use

mystic remnant
#

i already changed that to add --api to the command line arge =

#

it doesnt change anything

glad pagoda
#

did u solve it

gilded crane
#

where should I ask questions?

#

about stuff I made in stable diffusion*

wind heron
#

Hey all. Could someone point me in the direction to jnfo on how to run SD on multiple GPUs. I have lots that I want to utilize.

static snow
#

Hello everyone, excuse me, I'm trying to get a woman with 4 or 6 breasts, but no matter what prompt I use it doesn't give me that, any tips?

wind heron
#

Also need to be running SD 1.5 since the new 2.0 and up nolongwr have nudity.

static snow
#

Well that sucks !

wind heron
static snow
scenic flame
#

do you think it is necessary to understand deep learning deeply to use modern AI technologies such as stable diffusion to generate the pictures I want? or can everybody use it?

wind heron
raven thorn
static snow
#

I suppose I could train it? but I don't know a google colab that trains embeddings

signal oxide
#

hi! i have a question, i cant type in any creation channels, do i have to follow a process or something 😛

vast pollen
#

Does anybody know something like anythingv4.5 that is not a troll model or whatever?

#

like no purple splotches over the images?

#

ping me if you know

vapid tangle
#

when might we get an alternate for chat gpt? that thing is so fking censored its annoying to get it to do much anything at all

fast hamlet
#

It's probably gonna be a while - currently LLMs aren't exactly possible to run on consumer hardware, making the open source approach difficult

astral goblet
#

not really. you can open source it, but still have it require memory

vast pollen
#

Okay if you get purple splotches, you need to get something called vae pt

astral goblet
#

which is just pulling more of the orangemix side of the blend out and muting part of the anything v3. the original av3 is the best way to go. orangemix is already got av3 in iit, so blending it with av3 again is just, well, that's what gets the splotches

vast pollen
#

oh, okay

abstract pivot
#

Hey, does anyone know if there is a tool or way to rip the prompt info from generated PNG's(with the metadata stored) as a batch process? I have a lot of pics I want to get the txt files for but the only way i can think of is manually throwing them back in SD(A1111) and reprocessing the entire image with the export .txt file setting enabled again... Thanks

forest apex
#

@abstract pivot yes it is in the info field of the png's exif

abstract pivot
empty barn
forest apex
warm junco
#

Maybe Chat GPT can write you a script

forest apex
undone marsh
#

Anyone from the Stability.ai team here? I'm the founder of AI Films and would like to reach out someone from the team to know a bit more about DreamStudio Pro. We are 270 creators ready to create AI Films and have some questions regarding the launch of DreamStudio Pro. Wrote an email and got no response. My DMs are open. Appreciate

abstract pivot
sudden axle
#

listen

#

its okay

#

for CPAgrip and that stuff

#

its not that bad

#

10min max of work

#

it is for ads without pics

#

cpagrip*

#

it works like this

#

i post ad with my code

#

and if some one went on it

#

i can see

#

and if someone leavee email

#

i get money

#

bcz he entered the giveaway

abstract pivot
#

where are the mods at lol

sudden axle
#

(Affiliate: #1165229)

#

this number

sudden axle
#

it will be 8th ban

warm junco
#

Dude stop this scam stuff you wont get any money too

#

Cheap Internet tricks trapping kids into giveaways nobody ever will win xD

sudden axle
#

i cant be scam man
you just leave email

#

i cant get anything from you

warm junco
#

Yea and then they spam you with ADS and scam Mails and sell the mail to advertisers

elder cradle
#

Wow, scambots evolved

#

Ooh

warm junco
#

Nothing is free if you pay with your data

sudden axle
#

them

rancid pier
#

#anime

#

Why can’t I find the channel?

marble harbor
#

Hello guys can someone help me with stable diffusion?

elder cradle
#

It's now moved into the forum

astral pivot
#

what folder do i put PT files in?

coral shoal
#

I just have to say it.... There is way too much anime in the stable diffusion world. It's washing everything else out. maybe it should all be kept separate?

fervent thunder
merry prawn
#

Where do I generate stuff? Can't seem to find the channel

astral pivot
#

does anyone know similar sites to huggingface and civitai

fervent thunder
#

How do you turn a child picture into something like what they want it to look like?

For example my child draws a building, how do I then get SD to interpret that?
Do I use img2img?

fervent thunder
#

keep researching it

#

you'll get there (:

#

I came here for help not to be told to go look elsewhere, but thanks for responding.

fervent thunder
#

if you could be more polite I have even more in depth answers the exact answer and experience in it that no one spoon fed me

#

but thanks anyway

upbeat gorge
#

Hi Everyone!
With a friend we just release a new App where you can create Stable Diffusion Images :).
I really like to have some feedback so if u can test it i will be very happy!
https://play.google.com/store/apps/details?id=com.dreamai.app

For now is only available for Android because we are a team of 2 (dev and designer).

thanks!!

fervent thunder
upbeat gorge
#

SD 2.0

fervent thunder
#

nice, custom Lora or trained 2.0 ?

#

looks impressive

upbeat gorge
#

Trained 2.0

fervent thunder
#

thats cool mate nice work, really hope that it kicks off

upbeat gorge
#

It was just a POC to have some fun :), we dont spect to be billionare with this, but we have some fun hehe

fervent thunder
#

I'm still creating custom mixes and embedding for 1.5 but hope to move to 2.1 soon and start making stuff for that, need money to rent gpu time lol

raw peak
#

Any know of another ai art genrator with the name of something close to podai. It has music in it as well. I forgot its name.

fervent thunder
#

try to prune the model you are using to remove ema

#

then load that pruned model

twin mist
#

@fervent thunder hey it's not bad ur image, why u've blocked me LOL

warm junco
bleak matrix
#

Good morning, lovely peoples of the chat! How are we today?

fervent thunder
#

yes it is meant to be strange asf

#

btw 😄

atomic swallow
#

hello

torn hull
#

Hey guys, I have a question. How do you find a model by its ID number?

distant tulip
#

Hey guys, I'm currently writing my master thesis in robotic related topic. I need a stable diffusion framework with the possibility of plugging in a classifier for guidance. I appreciate any kind of help! 🙂 I wish all of you a stable day.

rose grove
#

midjourney and lexica have amazing pieces of work, but seems like a lot is based on the lighting. Any ideas how to imitate their lighting?

vast pollen
#

Is this vae thing some update to AUTOMATIC1111?

#

Do I have to download it again to unlock this?

warm junco
vast pollen
proud dawn
#

Requesting help, is it possible to run fine tuned SD (with my own images) trough google colab and which tutorial should i follow if so? Im a noob at this stuff so i could use help.

azure blade
#

Came back to use stable diffusion after a few weeks, now the dream bot channels are deleted? Wt happened? How to create images now

robust coral
#

Thank you @wild steppe, @karmic brook and everyone involved in setting up the Dreamer Communities! Great work! purplestar

rose grove
#

Can xformer be toggled on and off, and is it worth just using it always?

signal dome
#

I don't see emad anymore in chat

#

Is he really super busy now

warm junco
polar summit
#

HI sorry to interrupter i am trying to instal stable diffusion but keep coming across this error

rose grove
polar summit
# 
ERROR: No matching distribution found for torch==1.13.1+cu117```
#

my error

warm junco
warm junco
polar summit
# 
==============================================================================================================
INCOMPATIBLE PYTHON VERSION

This program is tested with 3.10.6 Python, but you have 3.11.2.
If you encounter an error with "RuntimeError: Couldn't install torch." message,
or any other error regarding unsuccessful package (library) installation,
please downgrade (or upgrade) to the latest version of 3.10 Python
and delete current Python and "venv" folder in WebUI's directory.

You can download 3.10 Python from here: https://www.python.org/downloads/release/python-3109/

Alternatively, use a binary release of WebUI: https://github.com/AUTOMATIC1111/stable-diffusion-webui/releases

Use --skip-python-version-check to suppress this warning.
==============================================================================================================
Python 3.11.2 (tags/v3.11.2:878ead1, Feb  7 2023, 16:38:35) [MSC v.1934 64 bit (AMD64)]
Commit hash: ea9bd9fc7409109adcd61b897abc2c8881161256
Installing torch and torchvision
Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/cu117
ERROR: Could not find a version that satisfies the requirement torch==1.13.1+cu117 (from versions: none)
ERROR: No matching distribution found for torch==1.13.1+cu117

[notice] A new release of pip available: 22.3.1 -> 23.0
[notice] To update, run: C:\Users\mithu\Desktop\stable2\stable-diffusion-webui\venv\Scripts\python.exe -m pip install --upgrade pip
Traceback (most recent call last):
  File "C:\Users\mithu\Desktop\stable2\stable-diffusion-webui\launch.py", line 360, in <module>
# 
  File "C:\Users\mithu\Desktop\stable2\stable-diffusion-webui\launch.py", line 269, in prepare_environment
    run(f'"{python}" -m {torch_command}', "Installing torch and torchvision", "Couldn't install torch", live=True)
  File "C:\Users\mithu\Desktop\stable2\stable-diffusion-webui\launch.py", line 89, in run
    raise RuntimeError(f"""{errdesc or 'Error running command'}.
RuntimeError: Couldn't install torch.
Command: "C:\Users\mithu\Desktop\stable2\stable-diffusion-webui\venv\Scripts\python.exe" -m pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 --extra-index-url https://download.pytorch.org/whl/cu117
Error code: 1
Press any key to continue . . .```
polar summit
warm junco
#

this is your problem -> This program is tested with 3.10.6 Python, but you have 3.11.2.

polar summit
#

oh does it have to be the exact version

warm junco
#

not the exact but a supported version so 3.10.6 or 3.10.9 will work

#

3.11 is unsupported

polar summit
#

ahh i see danke schon

rose grove
#

Haven't had time to install it yet

rustic osprey
#

Looking to have someone creating an SD based web app. How do I go about finding someone for that job?

warm junco
tame moth
rose grove
tame moth
#

Guys I need help, I need to change Spritesheet (animation of the gun) for game, but if I use as input I get some rubbish, how to change picture, to get the same object but in different style?

livid flint
#

where i can generate images?

upbeat birch
rustic osprey
potent latch
#

hello - is there a way to make diffusion utilize a non-nvidia gpu? i've seen some tutorials which try and demonstrate how to achieve such a state, however after following a few i get stuck on one step or another, and so i default back to using just my cpu, which takes ages

polar summit
#

@warm junco sorry for disturbing you

# 
No Python at '"C:\Users\blank\AppData\Local\Programs\Python\Python311\python.exe'
Press any key to continue . . .```
#

I install the right python version you told me even restarted

#

even check the path

warm junco
#

did you uninstalled 3.11?

polar summit
#

??

#

3.10.6

#

i installed

tame moth
warm junco
#

you can set the python path in the webui-user.bat

polar summit
tame moth
polar summit
#

thank you redas and CS10

nova lily
#

PSA: Despite wherever you install StableDiff, it may still be using your C Drive to house temp images, which is likely your SSD. You may already have 1000s of images written there already. I suggest changing that directory immediately. (setting is under watermark)

stable pawn
#

I have some trouble in Automatic111 with viewing tiled images, other have same trouble?

tame moth
heavy crater
#

the server is about stable diffusion?

#

im having an imense ammount of trouble with it

#

and i have absolute no idea of how anything works

wild steppe
heavy crater
#

aight

#

thanks!

dawn basalt
#

Anyone Here Have an electric Sitting/Standing Desk?

maiden crystal
#

@warm junco Hi. So i gotta reinstall SDwebui. and i noticed the Aitrepreneur his MEGA download files, says Nonema. But I try to download v2-1_768-nonema-pruned.yaml

It always says ema, but not "non-ema"? Im very confused whats goin on. even on MEGA it says nonema

#

update
Nevermind I found it. Just downloaded all as ZIP , and it was in htere.

#

wonder why he gives Non ema version in mega, if thats for training?

#

oh nevermind i gotta rename it.

warm junco
#

Good you got it working, the model files are also on huggingface

maiden crystal
#

like for auto launch etc

#

-- and
do you know good prompts to prevent the image from tiling? like making wallpaper patters?

shut owlBOT
#

"TakeRep" command returned an error: strconv.ParseInt: parsing "and\ndo": invalid syntax

warm junco
warm junco
maiden crystal
#

and it looks like a wallpaper or wrapping paper pattern

#

it will make many small images, in one image

#

and i want it to focus on one image, center canvas

warm junco
#

Ah you have to lower the resolution and use highres fix or sd upscale To upscale it

maiden crystal
warm junco
#

Not at pc rn sry

maiden crystal
#

np ok

#

is there a prompts channel i can ask about this ?

astral goblet
#

Anyone got any tips for testing trained embeds? just getting started on this stuff now and i ran it on training all night 100000 steps. saved embeds every 500. so i got a lot and all the example images look fairly well trained after 1500 steps. Curious to what a good xy grid approach would be

warm junco
#

Then you can say Embedding_xxx, and use xxx, 100, 300, 500, etc

astral goblet
#

i could probably go rougher, every 5000, then when i find one, look closer into that local range

warm junco
#

Yea good idea

astral goblet
#

i'd show some results but my friend is paying me for this and told me to keep it private, no posties lol

warm junco
#

Haha okay xD if its his face it makes sense

astral goblet
#

i'm pretty impressed with how well his face turned into an embedding though

ancient mango
#

I tried analog diffusion and dreamlike, anyone know another model that's realistic that could be used to generate pictures of stuff for like tinder.

#

social media in general

atomic jolt
#

what is the best height and width for portrait format

warm junco
#

512x768 is very good

atomic jolt
#

thx

marble pendant
#

My ass itches

#

Hi chat, btw

#

How is your day going?

primal nebula
#

hello, question, stable doesn't have a prompt book yet right? or a channel here like midjourney?

molten bough
proud dawn
#

Looking for tips on fine tuning stable diffusion. I have like 72 images, is 100 steps per image enough for good results?

fervent thunder
#

if you turn denoising up to 1.0 on img2img is it essentially like doing text2image where it doesnt reference the image at all? 🤔

stark scarab
maiden crystal
#

How do I make the picture less abstract? Im trying to prompt for food images, and drawn or painted styles, and Its always like jumbled abstract and dont look just like a cake, or a cup or a cone, or a ice cream etc

#

im using 1.5 and 4

fervent thunder
#

try img2img with a real cake image or something

#

i mean its not photo real maker just yet, 1.4, 1.5 will reveal the imperfections like any ai image when you study it

hardy junco
#

lora files ( .pt) can to be a virus?

nova lily
#

maybe dumb question, but would/does a HiRes fix image come out better than if you upscale after the fact

brisk bloom
#

if you hires fix with original seed and prompt with denoise aroun 0.5 it will be much better than a simple upscale

nova lily
#

@brisk bloom Any advice on which upscaler

brisk bloom
#

swinIR

nova lily
#

👍 will play around

shut osprey
#

hello

sterile gust
#

HAI

fathom imp
#

Hi

astral goblet
fair trellis
#

wow

#

what happend

#

to the animie channel

scarlet egret
#

Hi, can someone please tell me why the result while being generated looks colorful but gets more gray when finished? I tried out different VAE

astral goblet
wise flax
#

DAMN IT!
male genitals are disgusting in sd!

#

change it

scarlet egret
#

is it possible to create a work in progress of the images?

mystic plinth
#

Hey , can someone help me fix the log vml cpu not implemented for half
Runtime error ?

dire sun
#

anything worth to update lately with automatic 1111?

crimson veldt
#

hey

warm tiger
crimson veldt
#

anyone here knows when sd 3 is coming out? i remember emad saying there will be an SD3

fast sedge
#

Stability AI posted about something coming at 5pm GMT

#

Does anyone have any idea what it's going to be about?

astral goblet
#

machine learning maybe

nova lily
#

Portraits come out great, but wider ratios have terrible composite and unnatural. Any advice to make wider portraits look better?

warm junco
warm tiger
#

so what do you do with a vae that's just .pt?

warm tiger
molten bough
scarlet egret
#

Hi, can someone please tell me why the result while being generated looks colorful but gets more gray when finished? I tried out different VAE

warm junco
warm junco
warm tiger
#

so just the .pt ones go in the vae folder?

#

i could have sworn those were the .vae files

#

but this one says it's just .pt and says to put it with the model

warm junco
#

Where did you found it?

cedar crystal
#

what is a pruned model? and how do ppl do it

stable pawn
#

I am running the Automatic1111 WebUi and I start it through a batch file that have the line "set COMMANDLINE_ARGS=--no-half", but I have not found any good help for the Arguments, so; if I have a Nvidia GTX 3090 and a okay CPU and I maybe want to train a own LORA and in the future install some extensions, is that all I need or would any here suggest other ARGS to add?

warm junco
stable pawn
#

Ty, what do --autolunch enable, it sounds practical.

warm junco
#

It opens the webui in Browser automaticly after you run the .bat

stable pawn
#

Ty, that is neat.

warm junco
#

No problem, you shouldnt have any Problems with training with a 3090

patent orbit
#

hmmm

fervent thunder
vernal stratus
#

what happened to this server? Are there somewhere category galleries? It is really bad move to remove them in my opinion

#

ok ^^ I see now - new structure

young ivy
#

anyone know a good lora guide?

chilly valve
#

morning Coffee waow

uncut plinth
#

Is there a new model that can also write text or has it not been released yet?

stable pawn
#

@young ivy I bookmarked a YouTube video called "How To Do Stable Diffusion LORA Training By Using Web UI On Different Models - Tested SD 1.5, SD 2.1", it is 59 min long, I will soon watch it, it is 1 month so it may be a bit outdated, so I am not 100 % sure if it is good, but many tutorials and guides I seen have been short and often stress over important details.

young ivy
warm junco
stable pawn
#

Ty, neat.

#

I guess @warm junco link is a better alternative then the video I suggested, for it look like many have problem with Dreamboth just now, it is rather fast, but that is what many like, I am just a old man that like it slow 😄

hallow osprey
#

Hello, I need help pricing my finetuning service for game art for a gaming company.
How can I price it per model per game per hour? Thank you.

jade rover
#

My account got hacked help

muted mural
#

hey guys, is there a separate discord for deforum only? i cna't find it online

#

but it seems to exist as pointed out by a youtuber

dense dragon
#

hi

tidal bough
muted mural
#

@tidal bough i managed to find it, thanks anyways for helping

fervent thunder
#

hi

green sandal
#

Do you guys know what is the best way for someone who is not tech savvy and doesn't have an nvidia gpu to be able to train a model and generate art of a certain style? i mean honestly i would pay for a package that i could just click a simple download button if that was an option. i already have folders of reference images organized. but my laptop and i are not capable of handling automatic1111 or navigating github, huggingfac, etc.

#

i only know these keywords because when i as around on different message boards these are the answers i get. but i know ive tried and run into problems because i have an amd graphics card, not nvidia. and its not a powerful laptop anyway.

#

i understand you can rent a gpu remotely but im starting to feel cross-eyed when im looking at python, and i have to make all these folders and download so many programs.

vale marlin
#

hi

waxen swallow
fervent thunder
#

there is no anime channel anymore?

karmic brook
swift sluice
peak sphinx
#

Does someone know any good models for generating LOGO?

proper karma
#

I thought stability was beyond blockchain MLM garbage 😭

#

what's this about a deal with krikey

peak orchid
wicked patrol
#

cant wait !!!!! to see !! 👀👀👀

lunar surge
#

It's time!

abstract pivot
#

.....

#

Hello? Anyone? 😂

wide juniper
#

I see people using (bad quality, worst quality:1.4) all the tim. Why 1.4 and not 2 or 10?

glacial stag
#

Any specific place where they make the announcement?

wicked patrol
#

⏰⏰⏰

abstract pivot
#

Late as always

sour frigate
abstract pivot
#

Hah, tbf the announcement of the announcement was on the stabilityAI Twitter not the Stable Diffusion Twitter, what were we expecting. Cool news for stability anyway 👍

swift seal
#

Hello everyone. Is there a way to force the prompting to only have one model? It seems to be making two people alot when I only want one.

glacial stag
#

You have to use Hires.fix

#

For res higher than 512x512

fervent widget
#

where to generate images and how?

swift seal
glacial stag
#

Yes

#

Hires fix … fixes your problem, so you don’t have two people.

fervent widget
wintry stream
astral goblet
#

I've been watching ML and medicine for a long while. I really like the potential it has for doctors to accelerate diagnostics and getting people more custom treatments. Cool that stability is making it known theyr'e working this space

thick wharf
#

hey all! I'm the founder of MedARC... Glad to see some interest here! 🙂
Be sure to join the server, lots of opportunities coming for contribution and lots of great research!

proper karma
astral goblet
#

as a canadian, our doctors and infrastructure desperately need tools to aid their paperwork and amplify their effictiveness. i love what you're doing @thick wharf !

proper karma
#

and obviously a bunch of other stuff... I'm imagining protein generation research, gene identification..?

#

I haven't looked into what MedARC does

proper karma
wild steppe
thick wharf
glacial stag
thick wharf
glacial stag
#

The fMRI project sounds very interesting…

thin widget
#

What are people using for SD image management on Linux? As in, browsing images post-gen.

fiery wing
#

Hey guys I've recently updated my stuff and all my pictures are sorta low-saturation or washed out art style. any ideas on how to fix this?

tidal bough
#

Hey everyone! We’ve got a stage event featuring Stelfie the Time Traveller coming up at <t:1675969200:t>! He’ll be talking all about the workflow used in creating his time travel selfies; Head over to Center Stage once the event starts, and drop your questions in #1029055412764422214 ! kneadforspeed

wild steppe
native root
#

great talk - thanks for setting it up!

wintry stream
#

Thanks for taking the time to bring us along on a journey working through the process with you @dense solar!

wild steppe
#

Thanks for joining everyone ❤️

tidal bough
#

Absolutely mind blowing stuff - Really takes inpainting a whooole new level! catwhaaa

summer patrol
#

I love stelfie

tight python
#

That live was amazing!

#

I got notified right on time. Sorry if it's easy to find info, but are there going to be more stages like these?

fervent thunder
#

Can i ask questions in general?

#

well, I already did. but I would like a bit of help with using models from Civitai. when i copt prompts it generates beatuful images. but really different

#

wewll, not "really"

fervent thunder
#

what do steps do

fervent thunder
#

Guys, how do I start generating?

wintry stream
forest apex
#

hey is there an archive of the centerstage stream? I missed a lot of it

#

totally waiting for that vox video that shows his process

wintry stream
lost notch
#

Hey guys how long does it take to merge 2 checkpoints?

#

Mine webui is kinda stuck doing it

warm junco
#

It should take less than a minute

lost notch
#

weird... mine webui is stuck on loading chkpt B

#

when i tried to reload it got me disconnected

warm junco
#

How much RAM do you have ?

lost notch
#

16

#

and 8 in GPU

warm junco
#

Make sure you have enough RAM free

lost notch
#

i do

#

it wasn't using RAM at all

#

and when i tried to make the merging again it was added to queue ?

#

what queue 😅 🤣

warm junco
#

Maybe some step was stuck

#

Try restart and then only merge

lost notch
#

what option should I choose from "copy config from" ?

#

A,B or C; B ; C ; don't

#

maybe I choose wrong option here

neat dome
#

I read on a forum that running stable diffusion locally using a gddr6x card runs the risk of bricking the gpu is this true has anyone seen this ?

timid plank
#

so im new im still trying to get it to work
what does "To update, run: H:\AI\stable-diffusion-webui\venv\Scripts\python.exe -m pip install --upgrade pip" mean

#

and "no module 'xformers' proceeding without it"

astral goblet
#

it's just software interfacing to the driver

neat dome
ancient mango
#

idk if im not giving dreambooth enough pics to train it but every model i use gigachads tf outta the pics of me I give it lmao

spring tundra
stray tangle
#

is the bot gone

#

i dont even remember what its called

spring tundra
indigo matrix
#

Aonde eu crio as minhas imagens? não tenho o dreambot

#

Algum BR aqui pode me ajudar?

#

Kkkk

nocturne mesa
#

gm

fringe minnow
#

any tips on language to use designing poker chips?

mint matrix
#

What new SD models have popped up for the anime genre?

#

Its been a while since I popped back into this server

alpine bear
mint matrix
#

I see

alpine bear
hybrid frigate
#

has there been any confirmed cases of malicious pickle yet

stuck shoal
#

Malicious Pickles is a pretty good punk band name

fervent thunder
sour marsh
#

did the server name changed? (i was raidk now im d127c389)

vestal dew
alpine bear
young ivy
#

you can just go change it again I think

crisp verge
#

hi

past lantern
#

not sure if this is the right place but... I am using the automatic1111 web ui and I have no idea where to put my models I downloaded for it to see them. I downloaded a few models from civitai

warm junco
past lantern
#

do I have to do anything else because it does not seem to see them

#

I have a few safetensor files and .pt files and .ckpt files

warm junco
#

Yea i tought so, you need models, make sure they arent Loras (.pt or Safetensor with 1-300mb)

#

Only files over 2gb going into the models/Stable-diffusion folder

past lantern
#

some are small... but I have deliberate and dreamshaper and it cant see those

#

they are 3gb+

warm junco
#

Can you send a Screenshot of the commandline window?

subtle bolt
#

👋🏻 Hey guys
Are we still able to use prompts?

wide juniper
#

Is there a reason why LORAs seem to have become more popular recently?

brisk bloom
#

apparently they are really pain free to make and fast

wide juniper
#

I see. But why is it a recent thing?

brisk bloom
#

I guess someone worked out the settings and best workflow, plus Dreambooth updates have ruined a lot of peoples projects with bad results

#

if something is easier its natural to move towards that

wide juniper
#

Yes! Dreambooth hasn't been nice to me

#

Playing nice I mean

ornate pilot
#

Can someone summarise how we actually produce images now? Is there a software or app download? A website? Since dreambot is permanently gone? How does this work now?

lament steeple
#

I use a local install. I think there are instructions in the "Start Here" Channel

#

Looks like I have managed to paint myself into a corner. Through my usual process of generation, outpainting, inpainting, upscaling, cropping, regeneration, and compositing, I managed to create a picture so realistic that a certain web site now wants a model release from someone that doesn't exist :-s

warm junco
ornate pilot
#

Stable Foundation?

warm junco
ornate pilot
#

Ok downloading now, it’s a huge download for an app 😮

warm junco
#

Yea i think it comes with some models so you can generate images immedeality

vestal dew
#

this prompt matrix doesn't work

lament steeple
#

Prompt matrix pukes if your batch size is more than one. I ended up fighting with it for awhile today before I got it working

vestal dew
#

really

#

I had batch size 1

#

batch count was larger than 1

#

then it breaks batch count, so if I get rid of the extension, it only does one image even if batch count is set to multiple

lament steeple
#

That might mess it up too. Didn't pay too much attention earlier. It's primarily used for testing things out before you hit the final go

vestal dew
#

I did <elephant|lion> with batch count at 12. Got elephant over and over

lament steeple
#

ah!

#

Promp matrix isn't what you want. YOu're trying to mix an elephant with a lion?

vestal dew
#

the images were titled elephant lion...just combined the words

ornate pilot
lament steeple
#

try [elephant|lion]

vestal dew
#

No. I want elephant one time, lion the next

#

the readme says this: Use < > to create a group of alternate texts. Separate text options with |. Multiple groups and multiple options can be used. For example:

An input of a <corgi|cat> wearing <goggles|a hat>
Will output 4 prompts: a corgi wearing goggles, a corgi wearing a hat, a cat wearing goggles, a cat wearing a hat

lament steeple
#

Sounds more like you want to use dynamic prompts

vestal dew
#

i want to do exactly what the readme says

#

if I do <elephant|lion> in a <jungle|zoo> I expect elephant in a jungle, elephant in a zoo, lion in a jungle, lion in a zoo

warm junco
vestal dew
#

these things seem so simple, but they never are

warm junco
#

Dont use <>

vestal dew
#

the readme says to use it

warm junco
#

No it doesnt

vestal dew
fervent thunder
#

didnt see that coming.,...

wide juniper
#

Do Ti, LORA and Hyprnetworks require restart to work?

warm junco
vestal dew
#

I see

fervent thunder
#

I need help but feel stupid if I ask

#

😄

warm junco
#

Feel free to ask

muted mural
#

Hey dudes, is there a link or a page where I can find ALL currently released models for SD. I find models on reddit, youtube etc. but is there a list with all models currently released?

warm junco
#

No there isnt, most models are on Huggingface and Civitai.com

#

And there never be a list with all models, because the Community creates daily new ones

fervent thunder
#

lol

flint fiber
fervent thunder
#

import torch

torch.backends.cudnn.benchmark = True

muted mural
#

Thanks, Civit.AI is pretty close to what I'm asking.

fervent thunder
#

Idk where to put

#

thisa

vestal dew
#

I made a model that's a combination of all other models. It produces an image of a dot every time

warm junco
flint fiber
muted mural
fervent thunder
flint fiber
muted mural
fervent thunder
#

I feel so stupid I cant follow instructions

vestal dew
#

the dot my model produces is like the singularity at the center of the big bang. Everything is in there.

warm junco
#

Will not work with webuis

fervent thunder
#

ok thx

#

So I can forget about that ?

flint fiber
warm junco
fervent thunder
#

xformers I have

vestal dew
#

cant use xformers with my lowly GTX 970

fervent thunder
#

I dont want to bother you too much I search for a solution thx for help

flint fiber
warm junco
vestal dew
#

Now I realize with Xformers - the bad hands and feet are caused by the Decepticons

warm junco
#

Then restart

flint fiber
muted mural
#

Can anyone explain to me in brief what do the numbers in prompts mean (1.2 for example) and also how do brackets and multiple brackets impact the prompt as opposed to just typing it normally

warm junco
digital kiln
#

Wen

fervent thunder
hollow grove
#

How i can make a pic with the AI?

#

What is the channel for that?

elfin wharf
#

does anyone know any tools like moises i can just run on a computer to split different elements of a song?

hardy crater
#

Почему я не могу писать в каналы с чат-ботом? Как снять замок?

fervent thunder
#

How do you animate portraits so that the character looks to be breathing, blinking etc ? I would love to add some life to my fake people.

#

nightcafe studio, starry ai are 2 I use.

warm junco
#

Stablehorde

muted mural
#

How do I get celebrities to look like themselves. The only ones I managed to get to look perfect are Messi and Ronaldo(I'm doing footballers), other famous footballers such as Maradona, Mbappe etc. only look like their brothers at best. I tried inpainting with different models, strength etc. and it just doesnt want to make them look perfect.

astral goblet
#

@tidal bough ~~ edit ~~ cheers

astral goblet
fervent thunder
tidal bough
#

@twin willow If you're interested in collaborating with the community on a project, you might want to drop any ideas you have in the #1045770510153302086 channel! Please hold off on posting too many personal details though; You can always discuss things further in DMs later on with anyone you end up working with!

astral goblet
fervent thunder
astral goblet
fervent thunder
astral goblet
#

it'll certainly do super life like idle animations too. whatever source video you put into it. nvidia may have some toys for this sort of thing too

fervent thunder
fervent thunder
astral goblet
#

he may use older tools but i wouldn't doubt he'll get into thin plate splines

fervent thunder
astral goblet
#

yeah i'm not about to take computer lessons from a meth cook psychopath no thanks! (jk thanks)

fervent thunder
obsidian meadow
#

Hi all

#

It's a long time ago that there has been an update for A1111?
There are many bugs already...

#

Is there something going on?

#

can we expect any updates?

#

Or is this the end?

#

Hope not...:(

#

I am used to a couple updates a week..

fervent thunder
#

I think the guy behind it has outside commitments and real life responsibilities. Updates will come I am sure.

floral geode
#

Hello folks I am new here
I used stable diffusion in my local but If I want to include it in a react project to make an ai art generator what is the process?
I couldn't find the proper docs about that.
I know about the official API but that comes with a cost is there any way I can make an API for myself and host myself for free for a limited time.
It is for my college project

obsidian meadow
fervent thunder
obsidian meadow
#

Issues for a1111 are 1200+

#

I hope its not a one man show

real zodiac
#

the real issues are sadly burried amongst the 'guise hoe do i install pytouch' ones hmmm

astral goblet
#

when doing major work too, its better to not release it incrementally sometimes

floral geode
astral goblet
floral geode
#

yes anything for refernece

#

there are so many about openAI though

#

but the dreamstudio api I can't afford

floral geode
#

if you can link me with some projects, I can go check myself

astral goblet
floral geode
#

I have installed stable-diffusion in my local system

#

it's awesome

astral goblet
#

if you are running this on your own PC, just need to wire up the networking right and it's a hosted api

floral geode
#

but how can I access it with the DNS

astral goblet
#

networking has never been my strong suit. that's a whole other can of beans

floral geode
#

I need my pc to power it up right.

#

but if I want to have it present in college I want to have a deployed backend and frontend

#

that's where I am stuck at

astral goblet
#

with automatic1111 there's a command line switch, --listen. makes it host on the network, not just the loopback. once it's on the network it's just a matter of getting all the layers to port forward correctly.

#

or run some sort of vpn to expose your server

floral geode
floral geode
astral goblet
floral geode
#

yes that's where I am thinking of hosting it somewhere

#

but it's expensive elsewhere

astral goblet
#

sure is!

floral geode
#

where it would be on 24*7

#

yeah

astral goblet
#

you'll be better off with on demand hosting. shadowpc for cloud gaming, i imagine would work really well

floral geode
#

yes

#

but tell me one thing

astral goblet
#

starts at $40/m though

floral geode
#

that google colab where the ui is running can't I use that as an API

#

I don't have idea on that

#

just asking

#

or like huggingface

astral goblet
floral geode
#

ohh

#

Got it

#

got it

#

btw thanks for the reply

#

I was so confused

astral goblet
#

theres a lot to sort through

floral geode
#

also you mentioned

#

there are lot of projects

#

can you refer me something

#

actually I wanted to build something a smaller version of lexica

#

for a small college project

#

lexica.art

astral goblet
#

not really. i am lost in the deluge of information as well. don't really have any good bearings out here quite yet

floral geode
astral goblet
#

i'd suggest finding a small project and digging into how they wired up the webapp to the stable diffusion process

floral geode
#

yes, where can I find?

astral goblet
#

that's the only bit you'd need. i'd imagine you're already adept with making a webapp

floral geode
#

whatever project I saw there is not GitHub repo

floral geode
astral goblet
#

you'd understand your needs better than i. just find a small one on github or pick through that 6month old list. like i said, this is all so new that theres no real "documentation"

floral geode
#

I can make a frontend and backend work

floral geode
astral goblet
#

gotta get going. work beckons

floral geode
#

yes, thanks a lot

mossy prism
#

what's a diffesur I don't get it

fervent thunder
#

also have a 2.1 mix on my profile

azure gazelle
#

Does anyone have a fix for the new discord image clipping? It’s so hard to see generated content now.

mossy prism
#

I need help

#

how do I fix : RuntimeError: Cannot add middleware after an application has started

#

wrong chat

cedar glen
#

heyyyy is there anyone here with experience selling on OpenSea?

proud nova
#

Yes. Finally free space on Paperspace 😎

runic oar
cedar glen
#

Lol apparently

#

I am only trying to get into because I keep getting requests from people wanting to purchase my art as nfts but I would love to talk to someone who has experience with it

thick steeple
#

Hello does anyone nows if there is still some google colab notebooks that to train with dreambooth?

#

Those used to be very common a few months ago but most are very outdated

fervent thunder
proud nova
#

I installed it a while ago.

hardy elk
#

Hello!! Does anyone know how can i request a personalized art? I need a black french bulldog with a french pastry and or a chef hat

errant stump
#

Hello

dim tundra
#

When img2layers?

#

That would be an AWESOME use of the 30fps of the upcoming new model.

livid basin
#

I am new

grim crown
#

hi guys

vestal dew
lament steeple
fervent thunder
#

Man I wish there was a TI or a LORA for the Yoji Shinkawa style of art

#

That's the metal gear solid guy

lament steeple
fervent thunder
#

I don't think it'll do much but I will

#

Currently my computer is being repaired haha so I can't do anything

lament steeple
#

see "general-with-images" in about 5 seconds

#

The link I just posted is every artist known to stable diffusion with the year they died unless they are still alive and whether or not they are in the public domain

fervent thunder
#

Can't find it

lament steeple
#

I exported sql queries to csv files. Took me about 3 days to look up all the artists

fervent thunder
#

Thx

lament steeple
#

np

woeful elbow
#

can i download stable diffusion and use it on my computer?

trail ledge
#

yes, assuming you have a decent gpu

#

there are multiple UIs out there, such as Automatic1111 and InvokeAI

maiden crystal
#

runs from pc, and has browser interface

woeful elbow
#

really?

maiden crystal
#

if you search my name, i had asked about that recently, and all my comments and the replies given me should set you up with it

#

in general and tech support channels

woeful elbow
#

im having a error and it might be the webui

#

can you check?>

maiden crystal
#

im like you novice with this. so all i did was install python 3.10 first , and go from there.

#

what error say?

woeful elbow
#

ERROR: Could not find a version that satisfies the requirement torch (from versions: none)
ERROR: No matching distribution found for torch

maiden crystal
#

running it on ur pc is taxing, i will say that.

maiden crystal
woeful elbow
#

yes

maiden crystal
#

not 3.11

maiden crystal
#

follow that video

#

if you hit all the points in that video, if anything messes up ill try to help. but im novice

#

can install automatic111 from the cmd, from the folder, after u install the git thing

#

and use the address he got in the description

#

maybe be sure your fire wall not blocking something as that downloads etc.

ornate pilot
#

I need the latest iPad pro with M2 chip to run this drawthings.ai app properly…. Image generation takes so long and It keeps crashing my 2020 normal iPad

vestal dew
#

it runs fast on an M2?

nova lily
#

If you parenthesis a word that's within a multi word prompt does it have effect?

#

well after trying it I think yes-sish

maiden crystal
#

im still looking for a prompt glossary. is there one ?

trail ledge
#

not that I know of

vestal dew
surreal radish
#

Man Chat GPT is so useful. Told it to write a python script that searches through .txt files for a tag and then sort all the corresponding images into an output folder. Makes organizing for model training so much easier.

queen breach
#

Is it possible to pass the settings between several users?

frail cloud
#

I keep getting a runtime error when imma try to launch the webui

#

Been stuck on this for 2 hours I’m not much a tech person so I don’t understand how to fix it

static agate
#

guys any artist out there can help with an 3d Animation ? for more details dm me thanks

maiden crystal
# frail cloud I keep getting a runtime error when imma try to launch the webui

dont stress it. take a brake come back to it tomorrow or later. and maybe take in more videos, take note of what pc specs you have exactly. and try again if possible.

All else fails, can use the online AI art generators, like Mid Journey ( not free ) but much more advanced in terms of results, and easier to use. Works through their Discord server.

placid ore
trail ledge
#

I'm still using Automatic1111 and InvokeAI, so I haven't tried anything else

maiden crystal
placid ore
#

im still using an old SD lmao

#

checking which version

trail ledge
#

I just started playing with it. I like the idea of the unified canvas, but it seems that automatic1111 still has more bells and whistles

placid ore
#

i heard everything past 1.2 was messed up

#

neutered

#

so i was afraid to update

maiden crystal
#

but if its a older model, i think im using with A1111. unsure why the results are different.

#

[ at least it seems to be better most times ]

trail ledge
#

wish they listed the version on the website

white niche
#

Open datasets that are useful for SD, that aren’t laion. Does anyone know of any that are based on open data with captions etc? Either image based in a folder or Parquet. Just curious what exists that is useful to the community and isn’t targeted at huge companies etc.

static snow
#

CUDA out of memory. Tried to allocate 1.25 GiB (GPU 0; 12.00 GiB total capacity; 9.02 GiB already allocated; 0 bytes free; 10.10 GiB reserved in total by PyTorch)

#

I just bought a 3060 for 12gb and I get the same errors

#

what the heck

#

any tips and I'll love you forever

lament steeple
#

What's your batch size and resolution? Try lowering it

neat bronze
astral goblet
elder cradle
#

--medvram also?

astral goblet
#

not with 12 yoiu shouldn't need it

rugged perch
#

is there still updates or is stable diffusion dead

astral goblet
#

lol is it dead

elder cradle
#

LMAO, "SD is dead"

astral goblet
#

obvious troll is obvious

#

ask a legiitimate question if you want a real answer

elder cradle
#

No updates from Stability AI, just that we're not receiving any info on the new model

#

Asking whether it's dead is really weird, since the company is still fine

lament steeple
#

Opensource projects don't die

elder cradle
#

That's right, there are a crapton amount of models based on SD on the Internet

lament steeple
#

Ugh! I love it when SD signs my work for me. 😛

raw swallow
#

I really got banned from Unstable Diffusion for saying that the mods were uptight lol

#

can't believe they give modding privs to children these days

worn marlin
#

guys, how to use LoRA?

unreal cargo
#

If you're using Auto1111, you just pop it in the "lora" folder in stable-diffusion-webui -> models -> lora. Then there's a red book button on the right hand side under generate. You can find all your LoRAs there after you refresh the list.

#

Just click on the LoRA and it'll automatically add it to your prompt

bright lava
#

is the strength of your lora the number after the colon when it adds to your prompt?

#

its funny i was just trying to figure this out and here the conversation is exactly this.

severe zealot
#

hi peeps im trying to change my self and friends into anime characters but still resemble the likeness using the Anything v3 ckpt model, any tips or setting i can use to make it work better? not many videos on it

unreal cargo
unreal cargo
#

like through model merging alicatHm2

#

There's img2img, you just generate a picture you generated of yourself then Img2img it using Anything 3

#

You could train on a merge of Anything v3 and 1.5 and generate off that alicatHm2

severe zealot
unreal cargo
#

yep np alicatLove

placid elm
#

Hi, can anyone help me with fixing Cuda Memory Error?

#

using 12 GB RTX 3060

fervent thunder
#

hi, I have a problem, in any browser the connection is denied when starting the stable diffusion web version, how can I fix it? for some reason I can't send a photo in this chat, if someone agrees to help me, please write to me in the bos

placid elm
#

I am not able to solve my issue but let's see what I can help with

versed yoke
#

hey i am getting this error when i try to install stable difusion automatic 1111 ###############################################################
Python 3.9.12 (main, Apr 5 2022, 06:56:58)
[GCC 7.5.0]
Commit hash: ea9bd9fc7409109adcd61b897abc2c8881161256
Installing requirements for Web UI
Launching Web UI with arguments:
No module 'xformers'. Proceeding without it.
Loading weights [6ce0161689] from /home/boy21/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned-emaonly.safetensors
Creating model from config: /home/boy21/stable-diffusion-webui/configs/v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(0):
Model loaded in 6.4s (create model: 0.6s, apply weights to model: 2.4s, apply half(): 1.5s, load VAE: 1.3s, move model to device: 0.5s).
Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch().
Traceback (most recent call last):
File "/home/boy21/stable-diffusion-webui/launch.py", line 361, in <module>
start()
File "/home/boy21/stable-diffusion-webui/launch.py", line 356, in start
webui.webui()
File "/home/boy21/stable-diffusion-webui/webui.py", line 232, in webui
app.add_middleware(GZipMiddleware, minimum_size=1000)
File "/home/boy21/stable-diffusion-webui/venv/lib/python3.9/site-packages/starlette/applications.py", line 135, in add_middleware
raise RuntimeError("Cannot add middleware after an application has started")
RuntimeError: Cannot add middleware after an application has started

#

how can i fix this?

twin mist
#

u should use python 3.10.6 first of all

versed yoke
twin mist
#

ehm

#

use the installer 64 bit

#

unistall the other one

#

and tick "enable PATH"

versed yoke
#

i use conda

twin mist
#

conda install -n py310 python=3.10

versed yoke
#

strange it installed it but it is still using 3.9.12

#

The following packages will be UPDATED:

python 3.10.6-ha86cf86_0_cpython --> 3.10.9-he550d4f_0_cpython None

Proceed ([y]/n)? y

Downloading and Extracting Packages
python-3.10.9 | 24.7 MB | ################################################################################## | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
Retrieving notices: ...working... done
(stable) boy21@os:~/stable-diffusion-webui$ conda deactivate
(base) boy21@os:~/stable-diffusion-webui$ conda activate stable
(stable) boy21@os:~/stable-diffusion-webui$ bash webui.sh

################################################################
Install script for stable-diffusion + Web UI
Tested on Debian 11 (Bullseye)
################################################################

################################################################
Running on boy21 user
################################################################

################################################################
Repo already cloned, using it as install directory
################################################################

################################################################
Create and activate python venv
################################################################

################################################################
Launching launch.py...
################################################################
Python 3.9.12 (main, Apr 5 2022, 06:56:58)
[GCC 7.5.0]
Commit hash: ea9bd9fc7409109adcd61b897abc2c8881161256

twin mist
#

"To use python 3.10 or newer, you need to update to conda 4.11+. If you cannot update conda, you need to revert to python 3.9 or older."

versed yoke
#

my conda version is conda 22.9.0

twin mist
versed yoke
#

f

signal imp
#

where are the dreambot channels ?

versed yoke
#

but i still get errors

#

Launching launch.py...
################################################################
Python 3.10.6 (main, Nov 14 2022, 16:10:14) [GCC 11.3.0]
Commit hash: ea9bd9fc7409109adcd61b897abc2c8881161256
Installing requirements for Web UI
Launching Web UI with arguments:
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 1.16.0-unknown is an invalid version and will not be supported in a future release
warnings.warn(
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 0.1.43ubuntu1 is an invalid version and will not be supported in a future release
warnings.warn(
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 1.1build1 is an invalid version and will not be supported in a future release
warnings.warn(
No module 'xformers'. Proceeding without it.
Loading weights [6ce0161689] from /home/boy21/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned-emaonly.safetensors
Creating model from config: /home/boy21/stable-diffusion-webui/configs/v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(0):
Model loaded in 9.4s (load weights from disk: 0.2s, create model: 0.6s, apply weights to model: 5.0s, apply half(): 0.9s, load VAE: 1.9s, move model to device: 0.6s).
Running on local URL: http://127.0.0.1:7860

#

To create a public link, set share=True in launch().
Traceback (most recent call last):
File "/home/boy21/stable-diffusion-webui/launch.py", line 361, in <module>
start()
File "/home/boy21/stable-diffusion-webui/launch.py", line 356, in start
webui.webui()
File "/home/boy21/stable-diffusion-webui/webui.py", line 232, in webui
app.add_middleware(GZipMiddleware, minimum_size=1000)
File "/home/boy21/.local/lib/python3.10/site-packages/starlette/applications.py", line 135, in add_middleware
raise RuntimeError("Cannot add middleware after an application has started")
RuntimeError: Cannot add middleware after an application has started

twin mist
#

pip install --upgrade --user setuptools==58.3.0

versed yoke
#

this gieves me another error Collecting setuptools==58.3.0
Downloading setuptools-58.3.0-py3-none-any.whl (946 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 946.2/946.2 kB 8.7 MB/s eta 0:00:00
Installing collected packages: setuptools
Attempting uninstall: setuptools
Found existing installation: setuptools 62.1.0
Uninstalling setuptools-62.1.0:
Successfully uninstalled setuptools-62.1.0
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
spyder 5.1.5 requires pyqtwebengine<5.13, which is not installed.
conda-repo-cli 1.0.4 requires pathlib, which is not installed.
anaconda-project 0.10.2 requires ruamel-yaml, which is not installed.
spyder 5.1.5 requires pyqt5<5.13, but you have pyqt5 5.15.6 which is incompatible.

#

why do i get so manny errors

twin mist
#

no w8

#

pip install --upgrade --user setuptools==67.2.0

#

there was the old version in the command

versed yoke
#

hmm

#

i think i have fixed the other issues but this is the last error message To create a public link, set share=True in launch().
Traceback (most recent call last):
File "/home/boy21/stable-diffusion-webui/launch.py", line 361, in <module>
start()
File "/home/boy21/stable-diffusion-webui/launch.py", line 356, in start
webui.webui()
File "/home/boy21/stable-diffusion-webui/webui.py", line 232, in webui
app.add_middleware(GZipMiddleware, minimum_size=1000)
File "/home/boy21/.local/lib/python3.10/site-packages/starlette/applications.py", line 135, in add_middleware
raise RuntimeError("Cannot add middleware after an application has started")
RuntimeError: Cannot add middleware after an application has started

twin mist
#

pip install fastapi==0.90

#

pip install fastapi==0.89.1

#

try both

#

or pip install fastapi==0.90 --upgrade

warm junco
#

or add this line fastapi==0.90.0 in requirements_versions.txt

versed yoke
#

now i have even more erros

#

Launching Web UI with arguments:
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 1.16.0-unknown is an invalid version and will not be supported in a future release
warnings.warn(
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 0.1.43ubuntu1 is an invalid version and will not be supported in a future release
warnings.warn(
/usr/lib/python3/dist-packages/pkg_resources/init.py:116: PkgResourcesDeprecationWarning: 1.1build1 is an invalid version and will not be supported in a future release
warnings.warn(
No module 'xformers'. Proceeding without it.
Loading weights [6ce0161689] from /home/boy21/stable-diffusion-webui/models/Stable-diffusion/v1-5-pruned-emaonly.safetensors
Creating model from config: /home/boy21/stable-diffusion-webui/configs/v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying cross attention optimization (Doggettx).
Textual inversion embeddings loaded(0):
Model loaded in 3.6s (create model: 0.7s, apply weights to model: 0.8s, apply half(): 0.7s, load VAE: 0.9s, move model to device: 0.4s).
Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch().
Traceback (most recent call last):
File "/home/boy21/stable-diffusion-webui/launch.py", line 361, in <module>
start()
File "/home/boy21/stable-diffusion-webui/launch.py", line 356, in start
webui.webui()
File "/home/boy21/stable-diffusion-webui/webui.py", line 232, in webui
app.add_middleware(GZipMiddleware, minimum_size=1000)
File "/home/boy21/.local/lib/python3.10/site-packages/starlette/applications.py", line 135, in add_middleware
raise RuntimeError("Cannot add middleware after an application has started")
RuntimeError: Cannot add middleware after an application has started

twin mist
elder cradle
#

How can people be so dumb that they can go to whatever channels and then think they can generate images there

fervent thunder
#

Hi

#

pardon my ignorance, but I was looking for a public AI initiative - it seems stability AI is close to it, but is privately owned

#

am I right?

signal dome
#

Sd need to release there own language models soon since the competition for chat bots is at heat rn

dusty fossil
#

retiraram o chatdream

warm junco
signal dome
fervent thunder
warm junco
fast fox
fervent thunder
echo tulip
#

I have some issue using the imgtoimg system, because every output it give me monochrome colors like the image look quite saturated, anyone know why?

blissful flame
#

mercedes or bmw?

placid elm
fervent thunder
#

What about wujek cisco human network on stable diffusion ? or Michael

sullen abyss
#

hi , i was wondering how hard is it to create a model ?

junior pollen
#

Hi, I installed 2.1 version 768 on my PC but when I tried generating images in doesn't generate - instead if give me this error under the image space in the webui:

NansException: A tensor with all NaNs was produced in Unet. This could be either because there's not enough precision to represent the picture, or because your video card does not support half type. Try setting the "Upcast cross attention layer to float32" option in Settings > Stable Diffusion or using the --no-half commandline argument to fix this. Use --disable-nan-check commandline argument to disable this check.

#

how to fix this?

warm junco
junior pollen
#

RTX2070 with Max-Q Design

#

SD1.5 runs well

warm junco
#

Okay Do you have the right .yaml for 2.1 model?

junior pollen
#

yes

old carbon
#

Hello everyone. I am looking for a way to use SD (automatic 1111 webui version) with python or other scripting language to generate img2img with settings + prompts and for that image to be then saved as normally or to specific directory (whichever is possible). Is there a way to do this for example with python or command arguments in general? if so can someone point me to examples on how to execute this?

warm junco
#

Which 2.1 model do you tried?

junior pollen
#

2.1_768

#

do I need to add some lines in some files or whats the issue?

warm junco
#

Yes you have to edit you webui-user.bat and add --no-half behind Commandline_ARGS=

fresh meadow
#

hello fellow avocado chairs

old carbon
#

that sounds good, but how do i actually use it. Have a source with examples on how to run such prompts based on UI settings and others?

junior pollen
#

so "set COMMANDLINE_ARGS= --no-half" like this ?

#

the line after the = is empty

old carbon
#

ah i guess it wasn't for me 😛

junior pollen
#

now close the .bat and re-run it?

warm junco
obtuse breach
#

They're looking for a python API for auto1111

old carbon
#

@warm junco it's quite simple - i want to say in python 'run sd with arguments with settings with prompt with img and save it to directory' done

#

just not sure if that exists or if anyone knows how

warm junco
#

Ahh okay

obtuse breach
#

They even use Python in the example

old carbon
#

that does look good. thanks will have a look - it's always funny when the wiki with api actually is useful. doesn't happen often and I am pretty new to SD in general so my bad i guess ;P

#

thanks

obtuse breach
#

No prob

warm junco
#

@junior pollen you may also want to use xformers edit the bat again and add --xformers

junior pollen
#

what it does?

warm junco
#

It then uses less vram for generation so you can get higher resolutions and it improves the performance a bit

junior pollen
#

on which line do I add the --xformers ?

obtuse breach
#

Except for me where it says this version of Python doesn't support xformers FeelsBadMan

junior pollen
#

@obtuse breach which version do you use?

#

I have 3.10.6

warm junco
obtuse breach
#

3.9.13 apparently

warm junco
#

Autolaunch for auto startup in browser

sullen abyss
#

Is there tutorial on how to make models ?

junior pollen
#

Im getting some weird results with this model

#

can I show you?

warm junco
fervent thunder
#

--no-half will help

fervent thunder
#

I use

#

set COMMANDLINE_ARGS=--no-half-vae --xformers --opt-split-attention --medvram

#

on 8gb

warm junco
fervent thunder
#

I also work with 2.1 recently

warm junco
#

It just fills the vram to max and takes longer

warm junco
fervent thunder
#

had random issues on commit a few days ago might be fixed now but i had to use no half vae

warm junco
#

Okay yea some people need the half vae thing

#

But opt split is not nesesary

fervent thunder
#

Where's the room for crying? I'm so tired of fixing fingers , fire, smoke and faces...

fresh meadow
#

I'm having fun with training LoRA

obtuse breach
#

Is there a LoRA guide? I keep hearing the term, guessing it's like textual inversion?

fresh meadow
#

no, it's more like dreambooth and/or full finetuning in compressed form

#

in 10 minutes I created a 4mb file that works

#

just search the net for "rentry lora training", there are two good guides

warm junco
obtuse breach
vocal ferry
obtuse breach
fresh meadow
obtuse breach
#

Oh, I run automatic1111 locally

fresh meadow
#

this repo has also the required scripts

warm junco
obtuse breach
#

Ty

fresh meadow
#

I couldn't get into training DreamBooth because it creates 2GB files

#

LoRA is much simpler

quartz estuary
#

Hello, I'm relatively new to using invoke. Does anyone know a good guide for putting .pt or embetting into invoke? I'm not sure at all how to go about it.

tight bison
vestal dew
tight bison