🤝｜tech-support

[untold flame]

this has happened with dozens of files

[woeful pendant]

yeah, python is at fault

I use the one in miniconda myself

[woeful pendant]

def _join_cuda_home(*paths) -> str:
"""
Join paths with CUDA_HOME, or raises an error if it CUDA_HOME is not set.

This is basically a lazy way of raising an error for missing $CUDA_HOME
only once we need to get any CUDA-specific path.
"""
if CUDA_HOME is None:
    raise OSError('CUDA_HOME environment variable is not set. '
                  'Please set it to your CUDA install root.')
return os.path.join(CUDA_HOME, *paths)

is this windows aware or one of those linux only?

$CUDA_HOME is a linux thing while windows is %CUDA_HOME%

[untold flame]

It should be windows aware. I have had this same issue before with lots of installers

[woeful pendant]

def _find_cuda_home() -> Optional[str]:
"""Find the CUDA install path."""
# Guess #1
cuda_home = os.environ.get('CUDA_HOME') or os.environ.get('CUDA_PATH')
if cuda_home is None:
# Guess #2
try:
which = 'where' if IS_WINDOWS else 'which'
with open(os.devnull, 'w') as devnull:
nvcc = subprocess.check_output([which, 'nvcc'],
stderr=devnull).decode(SUBPROCESS_DECODE_ARGS).rstrip('\r\n')
cuda_home = os.path.dirname(os.path.dirname(nvcc))
except Exception:
# Guess #3
if IS_WINDOWS:
cuda_homes = glob.glob(
'C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v.*')
if len(cuda_homes) == 0:
cuda_home = ''
else:
cuda_home = cuda_homes[0]
else:
cuda_home = '/usr/local/cuda'
if not os.path.exists(cuda_home):
cuda_home = None
if cuda_home and not torch.cuda.is_available():
print(f"No CUDA runtime is found, using CUDA_HOME='{cuda_home}'",
file=sys.stderr)
return cuda_home

well, that looks for it 3 times

fails all 3

[untold flame]

I even manually went in and wntered CUDA_HOME = "the install"

in every single location where it pops up. it worked, until it tried to use CUDA and said CUDA is not installed on my system. Couldn't find a way around that one

Its like, for whatever reason it cannot see

Same with the packages saying they can't see the .dll, and then point me directly to it

[woeful pendant]

personally, I would start over. remove cuda, python, and nvcc

[untold flame]

I have tried that before, but at this point I will try it again

[woeful pendant]

well, don't just do it willy nilly

use miniconda for python

I will help you

[untold flame]

I'll follow anything you deem necessary. You have been an immense help for me in the past

[woeful pendant]

screw the python itself

let me get you the link

[untold flame]

I... Know how to do that, but I have programs that explcitly say "DO NOT RUN IN CONDA/MINICONDA"

[woeful pendant]

you aren't

[untold flame]

Ohhhhh, wait. I see what you mean I think

Its only using it to pull python from?

[woeful pendant]

pretty much but it is conda if needed too

[untold flame]

Ok... ok... I can work with that

[woeful pendant]

delete python

uninstall it

[untold flame]

on it

brb

alright back, almost done

[woeful pendant]

now uninstall cuda and the cudatoolkit

then reboot

[untold flame]

How do I go about making sure I properly uninstall both of those? I was looking into that and found no answers

[woeful pendant]

just uninstal via windows uninstal

[untold flame]

oh my

[woeful pendant]

omg, no wonder

[untold flame]

from all my attempts

[woeful pendant]

you blew its brains out

uninstall all of those

[untold flame]

I have had this issue from well before. Since the first day I got this PC back in Q3 2023

this... might take a moment

[woeful pendant]

well, there is a trick to this and just clicking install all can zap you like this

[untold flame]

this is painful 😅

[woeful pendant]

heheh, yeah that is a lot of dung

[untold flame]

going much faster now

almost done

[woeful pendant]

k

when done reboot

[untold flame]

anything special I need to do to uninstall nvcc?

[woeful pendant]

nope, just control panel

[untold flame]

should I remove these as well?

[woeful pendant]

poof

[untold flame]

I saw these get installed with cuda

on it

[woeful pendant]

I noticed you did an install all so yeah, issues

[untold flame]

all the error "help" i saw for my issues insisted I needed to

[woeful pendant]

then they well, nvm

[untold flame]

anything else?

[woeful pendant]

looks better

reboot

[untold flame]

ok, brb

back @woeful pendant

[woeful pendant]

ok

this is python 3.10.14 https://repo.anaconda.com/miniconda/Miniconda3-py310_24.3.0-0-Windows-x86_64.exe install it then we need to do some other stuff

[untold flame]

Alright

[woeful pendant]

install it

[untold flame]

installed

[woeful pendant]

now close any command prompts open

then open a new one

[untold flame]

admin or base?

[woeful pendant]

base

[untold flame]

ok

[woeful pendant]

in it type "python --version" without the quotes

[woeful pendant]

okay, not just me. here comes the fun part

[untold flame]

🙃

I'm in this deep lol

[woeful pendant]

you need to get into your system environment variables

[untold flame]

ok

I am there

[woeful pendant]

these three need to be added at the top

[untold flame]

for what Var?

[woeful pendant]

path

system not user

[untold flame]

ok, this i show it is currently

[untold flame]

so those three at the top?

[woeful pendant]

yes

[untold flame]

ok

[woeful pendant]

I have used all my space, it is crazy

[untold flame]

[woeful pendant]

yep

now apply and close

close command prompt and open a new one

[untold flame]

ok

[woeful pendant]

python --version

[untold flame]

:D

[woeful pendant]

there ya go

[untold flame]

A good time indeed

[woeful pendant]

now, do you still have the 12.1 toolkit? 11.8 is needed too but for now 12.1

[untold flame]

I have 12.2 as a lot of tools I use are asking for that, but I can do 12.1

[woeful pendant]

https://developer.nvidia.com/cuda-12-1-0-download-archive

12.2 is problematic, stay away

[untold flame]

downloading

[woeful pendant]

If this doesn't fix it then nothing will as it is not a system issue

now when we install this let me show you want to install

*what

[untold flame]

alright, installing. Custom install?

[woeful pendant]

yes

[untold flame]

alrighty

While its loading, wanna just say thank you so much man. You have always been a massive help

Truly appreciate everything you have helped me with and all the BS we got up to in the early days

even if they were fruitless... VoltaML lmao

alright

[woeful pendant]

uncheck all of those

[untold flame]

nuked

[woeful pendant]

uncheck cuda is lag from hell

[untold flame]

worked fine haha

[woeful pendant]

never does for me. now check the + next to cuda

only those

[untold flame]

ok, those 4?

[woeful pendant]

yes

[untold flame]

confirmation

[woeful pendant]

confirmed

[untold flame]

awesome

[woeful pendant]

then we must check something after you reboot

[untold flame]

sounds good

[woeful pendant]

hard part is over

[untold flame]

honestly... a cakewalk compared to the 10's of hours of BS I have put into getting around this error, and the stress it has given me lmao

ok, installed, restarting

[woeful pendant]

ok

[untold flame]

back

[woeful pendant]

k

now go back into the environment variables for the system

[untold flame]

I am there

[woeful pendant]

very top has?

might be lower

[untold flame]

for what? What var is that?

[woeful pendant]

path

always path

[untold flame]

top!

[woeful pendant]

perfect

[untold flame]

[woeful pendant]

now come out and is CUDA_HOME set because I don't have that myself

[untold flame]

it is not, but CUDA_PATH is, which is what it also looks for

[woeful pendant]

bingo

now open a command prompt

basew

nvcc --version

[woeful pendant]

perfect

this is as much as I can do

[untold flame]

ok! So not I try out what I was doing?

[woeful pendant]

see

[untold flame]

I will re-clone it all, of course

alright, need just a bit to get it ready

[woeful pendant]

k

[untold flame]

the moment of truth

my stomach hurts

[woeful pendant]

fingers crossed

if it doesn't nothing else can be done

[untold flame]

https://tenor.com/view/wanna-kill-myself-the-office-michael-scott-steve-carell-frustrated-gif-13414030

I don't fucking know anymore

[woeful pendant]

I highly suspect the program

[untold flame]

this has happened across soooo many though...

[woeful pendant]

you have the same setup as I do

[untold flame]

let me try another problem program

[woeful pendant]

let's try something

[untold flame]

let me try this other problem program real quick

[woeful pendant]

well, this puzzles me

you get that but you do not get No CUDA runtime is found, using CUDA_HOME

so it sees your cuda

[untold flame]

Lets see if my manual oobabooga install still reports missing DLL for exl2

thats the worst one I have had. DOZENS of times

every time I have fixed it it has been different

same deal. Can't find, points right to it

[woeful pendant]

I hate ooba it was crapola

so many bugs

[untold flame]

Yeah, agreed. I am on tabby now

[woeful pendant]

tabby?

[untold flame]

tabby has been good (been on it for like 3 days)

[untold flame]

extension for silly tavern that has really good EXL2 support. Allows you to load and generate with models right in silly tavern

[woeful pendant]

oh

go to that file you sent me and lets debug a bit

[untold flame]

it ALSO gave me the can't find exl2 dll error

I circumvented the check and it worked great

what file did we say it was again? I closed the cm,d

oh, nevermind

found it

[untold flame]

gotcha

[woeful pendant]

if we don't see made it to here there is an issue

[untold flame]

it did not make it

[woeful pendant]

it never printed that?

[untold flame]

it did not

[woeful pendant]

then there is the problem the program is bugged because it never went there to check for the path

[untold flame]

oh!!!!!

The no DLL error is gone!

Ooba worked

That has been the BIGGEST error I have had

[woeful pendant]

cool beans

[untold flame]

nevermind

I seriously wanna fucking kill myself lmao

how can this be so fucking broken

[woeful pendant]

that is ooba not windows

[untold flame]

its not ooba, trust me. I have had this error dozens of times with dozens of programs and dozens of DLL's

[woeful pendant]

as to the other issue if it never printed that then it never checked

the error is saying the dll does not exist

[untold flame]

it does tho. Like it literally points to it

just like how it can point to CUDA_HOME

[woeful pendant]

well, let's see what this does

alright I can guarantee this program is broken

it never finds it to set it

that is what the print was for

oh, wait

this says something bad

CUDA_HOME = _find_cuda_home() if torch.cuda._is_compiled() else None

the reason it never printed is because it says you do not have torch compiled

so it never goes to the routine

[untold flame]

that makes no sense... I have torch installed

[woeful pendant]

is it installed in its venv

[untold flame]

yes

[woeful pendant]

try this

[woeful pendant]

that matters not

get into that folder and cmd to it

[untold flame]

which folder? where those are installed?

[woeful pendant]

then venv\scripts\activate

[untold flame]

I am in an activated venv

[woeful pendant]

the one with the program's venv

[untold flame]

I am already in the activated venv

[woeful pendant]

good

sec

venv activated type python

import torch

[untold flame]

sorry, how do I type that exactly?

[woeful pendant]

(venv) python

[untold flame]

friend is saying to run this

python -c 'import torch; torch.cuda.is_compiled()'

[woeful pendant]

/>/>/>

I hate discord

[untold flame]

I do too, it can be such a pain

[woeful pendant]

>>>

that is what you will see in python

[untold flame]

ohhh, ok

I have it

[untold flame]

thats what it returns

[woeful pendant]

then

[untold flame]

....

[woeful pendant]

bad

bad torch

[untold flame]

😡

[woeful pendant]

thought so

your torch is cpu not gpu

[untold flame]

https://tenor.com/view/wtf-cat-grrrr-grr-grr-meme-gif-12171908424805132741

[woeful pendant]

hence, not compiled

[untold flame]

Why the HELL does it install that >:C

[woeful pendant]

you tell me cause they all do

[woeful pendant]

I suggest 2.1.2

[untold flame]

this requires 2.2.0 or greater

[woeful pendant]

then it is trash

[untold flame]

Trash? why?

[woeful pendant]

joking

here

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

type that from within the venv

[untold flame]

installing

[woeful pendant]

that is the gpu torch

2.3

[untold flame]

I am gonna blow up the dev lmao

||in minecraft||

[woeful pendant]

they all do it and look out if you install xformers

[untold flame]

Welp. I know to look out for this again in the future

[woeful pendant]

xformers will always remove yours and install a cpu version so needs --no-dependancies

now go back into python and check

[untold flame]

ITS BUILDING

I am gonna BLOW HIM UP >:C

[woeful pendant]

lol

now the dll exists?

[untold flame]

this is a different program, but I am moving the wheel to the problem one as we speak

[woeful pendant]

k

to install the wheel move it over, activate its venv and pip install the name of the wheel.whl

it will local install

[untold flame]

gotcha!

brb

[woeful pendant]

k

[untold flame]

ok so @woeful pendant

[woeful pendant]

?

[untold flame]

Having the same issue with yet ANOTHER

moved the wheel

wheel shows as installed

its in the pip freeze

I start it, it says its not installed

[woeful pendant]

what happens if you are venv activated and pip install it (regardless what freeze says)?

[untold flame]

lets see

wait...

it says

I deleted my python installs, what the FUCK

[woeful pendant]

now, delete the venv

get out of it and delete it

reinstall so it rebuilds the venv

[untold flame]

So apparently these are in my roaming

[woeful pendant]

remove them all

[untold flame]

windows says I have no python installed :/

nuking

ok, gone

[woeful pendant]

your python is in minconda and we know you had it as I had you cli python --version and it said 3.10.14

[untold flame]

yeah

making a new venv

going a little slow, sorry 😅

[woeful pendant]

no worries

a lot is in there

with as much change as we did always best to zap venv

no guarantees this will ever work as I had so many bugs in ooba and they were not helpful so I deleted it and lived free.

[untold flame]

I am taking care to make sure I install the right torch and copy the wheel over right as well

[untold flame]

I am not worried about it working, I am just trying to have the solace of knowing there is a solution in case I run into it again

Ok... here we go...

here's hoping

@woeful pendantI'm gonna crryyyy

it workedddddd

Happy tearssss

[woeful pendant]

YAY

might want to install that flash attention it mentions now

[fair oxide]

It keep failed to download? 😦

[untold flame]

Yeah, this was with no wheels. the minimum to open the Ui to make sure it didn't interfere with my exl2 wheel

[woeful pendant]

aye

[untold flame]

I am not actually gonna use ooba, but this is a proof face that building my own wheel (which I can do now) can fix the dll not found bug

I am so happy I could cry

you are such a fucking bro, man

Thank you sooooo much for the help. It means the world to me

[woeful pendant]

Nah, just been threough the washload a few too many times and hung out to dry twice as many times.

[ornate elk]

Hey, what's your GPU?

[woeful pendant]

You are welcome.

[ornate elk]

Restart the PC and try again

[untold flame]

💖

[fair oxide]

Still failed...

RPC failed

[ornate elk]

If your from Asia you may need a VPN for the download/installation part

Also make sure the PC date/time setting is correct

[fair oxide]

What's the best VPN for that? And the PC date/time is correct tho

[ornate elk]

Proton VPN should be good, has a free version

[wary maple]

nothing in the drop down list even i just created an embedding file. Any idea?

[cursive cobalt]

I has a few questions, all related to each other.

So, if I have a Lora, and a wildcard extension.
Could I theoretically take ALL of the prompts, from let's say... An armor pack, plop that into "ArmorPack1.txt", do a ArmorPack1, and have that randomly cycle through random combos?

If I wanted to do a batch over night?

__ ArmorPack1 __

Ffs

You know what the prompt is.

__ TXT __

That works I guess

[cursive cobalt]

Did u press refresh?

[crude galleon]

ive been trying to install stable diffusion/automatic 1111 andI've been recieving this error: error: RPC failed; curl 92 HTTP/2 stream 5 was not closed cleanly: CANCEL (err 8)
error: 2442 bytes of body are still expected
fetch-pack: unexpected disconnect while reading sideband packet
fatal: early EOF
fatal: fetch-pack: invalid index-pack output

does anyone know how to fix this

[wary maple]

yes i did

[cursive cobalt]

Then the other noob (Me) is out of ideas

[scarlet zinc]

hello there i have one question about the control API i using sketch to image AI can anyone help me out

[fair oxide]

Already used it and get US server but still failed

[fair oxide]

I get the same error even though I already use VPN

[crude galleon]

im using a vpn too, i tried turning off my vpn and it didnt work

i restarted my pc

didnt work either

[fair oxide]

Yeah, I did restart too but still error

[crude galleon]

my brother told me that its probably a network error but we both dont know how to fix it

i was fine before this

[fair oxide]

My network is fine lol

I found this when googling, maybe this will help us?
https://stable-diffusion-art.com/install-windows/

[ornate elk]

You can also try change the DNS server

Also use cmd instead of the terminal application

[fair oxide]

Alright, thanks. It's work but when I checked my GPU Vram, I didn't see dedicated gpu, only memory? So if it 11GB+, I don't have to add anything?

[ornate elk]

Click on GPU0 on the left side

Thats the dedicated GPU

[fair oxide]

This?

[ornate elk]

It has 4gb vram means --xformers --medvram --no-half-vae

[fair oxide]

Oooh, so it's written like that

What about this part?

GTX 16XX cards additional need: --no-half instead of --no-half-vae

[ornate elk]

You have an RTX card, so that doesn't apply to you

[fair oxide]

Oh okay, thanks.

[ornate elk]

No problem

[fair oxide]

Like this?

[ornate elk]

Yep

[fair oxide]

When I tried to running it, I got this message

[ornate elk]

That error appears mostly when an third party antivirus Programm is installed and blocked the Python install

Check if you have McAffe or AVG, Avast installed

McAffee is mostly preinstalled on many Laptops

[fair oxide]

Okay, do I have to use VPN when running the webui-user.bat?

[ornate elk]

First try without

If you get connection errors then yes

[fair oxide]

Immediately errors lol, alright I'll try with vpn

Both the same

[ornate elk]

What's the error?

[fair oxide]

This

[ornate elk]

Have you checked the PC for antivirus software?

Also try uninstall python and then install it again while the VPN is on

[fair oxide]

I have already turned them off

[ornate elk]

You then need to reinstall python while they are off

[fair oxide]

So, when uninstall and re-install, I have to turn the VPN on? Or just when re-install?

Alright, I already re-install it and the webui-user.bat is worked

[ornate elk]

Perfect

[fair oxide]

Now I have to waiting until it done. Hope no errors.

[spiral remnant]

Anyone having any issues with Controlnet and A1111? I hadn't used it for a long time, reinstalled A1111 from scratch 3 times, and that's with different versions of A1111 and the extension to try and see if something borked. But I'm just getting this error to say I haven't selected a model when I have (and mentions it loaded above the error). Have also tried SD1.5 and SDXL ControlNet models.. Am I doing something dense? I don't think I'm doing anything differently than a few months back

[sinful needle]

Hey mr wizard. Can you help me with something? I did a factory reset and now I'm starting from scratch. I installed Python, but it's telling me I don't have it when I start webui.

No Python at '"C:\Users\place\AppData\Local\Programs\Python\Python310\python.exe'
Press any key to continue . . .

[ornate elk]

Hey, delete the venv folder. Then install python 3.10.11 and select "add python to path"

[ornate elk]

How does your txt2img settings look like?

[fair oxide]

Do I have to wait until the cmd close by itself?

[fierce hollow]

Guys, I've been working with SD for about 3-4 months with images. I tried to make a video, it turned out terribly. Does anyone have a desire to become my curator and teach me how to make realistic videos? Unfortunately, I'm from Russia and I won't be able to pay you for this, but I think we'll both get the experience from this.

[spiral remnant]

Kept it nice and simple, base sdxl checkpoint too

[wet flame]

is this the right place to talk about the training of loras? how many images to use for training loras locally? I got lot of images

is 2k to 3k images good enough?

for a lora?

[sinful needle]

On A1111 startup

venv "D:\Ai\Automatic1111\Sourcefiles\venv\Scripts\Python.exe"
Python 3.10.0 (tags/v3.10.0:b494f59, Oct 4 2021, 19:00:18) [MSC v.1929 64 bit (AMD64)]
Version: 1.9.3
Commit hash: <none>
Couldn't determine assets's hash: 6f7db241d2f8ba7457bac5ca9753331f0c266917, attempting autofix...
Fetching all contents for assets
'"git"' is not recognized as an internal or external command,
operable program or batch file.
Traceback (most recent call last):
File "D:\Ai\Automatic1111\Sourcefiles\launch.py", line 48, in <module>
main()
File "D:\Ai\Automatic1111\Sourcefiles\launch.py", line 39, in main
prepare_environment()
File "D:\Ai\Automatic1111\Sourcefiles\modules\launch_utils.py", line 410, in prepare_environment
git_clone(assets_repo, repo_dir('stable-diffusion-webui-assets'), "assets", assets_commit_hash)
File "D:\Ai\Automatic1111\Sourcefiles\modules\launch_utils.py", line 177, in git_clone
current_hash = run_git(dir, name, 'rev-parse HEAD', None, f"Couldn't determine {name}'s hash: {commithash}", live=False).strip()
File "D:\Ai\Automatic1111\Sourcefiles\modules\launch_utils.py", line 165, in run_git
git_fix_workspace(dir, name)
File "D:\Ai\Automatic1111\Sourcefiles\modules\launch_utils.py", line 152, in git_fix_workspace
run(f'"{git}" -C "{dir}" fetch --refetch --no-auto-gc', f"Fetching all contents for {name}", f"Couldn't fetch {name}", live=True)
File "D:\Ai\Automatic1111\Sourcefiles\modules\launch_utils.py", line 115, in run
raise RuntimeError("\n".join(error_bits))
RuntimeError: Couldn't fetch assets.
Command: "git" -C "D:\Ai\Automatic1111\Sourcefiles\repositories\stable-diffusion-webui-assets" fetch --refetch --no-auto-gc
Error code: 1
Press any key to continue . . .

[wet flame]

yeah m getting some weird errors too for some on my desktop none of the UI is working properly

[hardy rapids]

trying to use supir on confyui with controlnet so i can up the hallucination while keeping similar composition not sure how to connect it. i did a LatentAdd node of the canny + the supir latent and connected to the supir sampler but getting this error.
"!!! Exception during processing!!! Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!
Traceback (most recent call last):
File "C:\Users\Tuan Do\Desktop\ComfyUI - TEST\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Tuan Do\Desktop\ComfyUI - TEST\ComfyUI\execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Tuan Do\Desktop\ComfyUI - TEST\ComfyUI\execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Tuan Do\Desktop\ComfyUI - TEST\ComfyUI\comfy_extras\nodes_latent.py", line 27, in op
samples_out["samples"] = s1 + s2

im able to do it on a regular img2img workflow with controlnet but not sure how to connect it with supir.

[wet flame]

ok so i found out the problem my driver is not compatible with cuda 11.x

it needs 12.x

and with torch 12 support instead fo 11

[ornate elk]

the webui will open in your browser when its ready

[sinful needle]

[ornate elk]

why didnt you git cloned the repo?

[sinful needle]

i... did

[ornate elk]

there seems to be an issue with the model

[sinful needle]

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

thats what i did

[ornate elk]

can you show me a new screenshot right now`?

[sinful needle]

of what

[ornate elk]

when starting up

[sinful needle]

so restart and paste info again?

[ornate elk]

okay git commit looks good but
wrong python version

only 3.10.6 up to 3.10.11 are supported

[sinful needle]

i downloaded the one u told me to. crap. gotta find it again

[ornate elk]

i dont tell anyone to download 3.10.0

[sinful needle]

oh

[ornate elk]

my guide in the pinned tabs links to 3.10.11

[wet flame]

its usually 3.10.6 to that

[sinful needle]

alrighty lemme find the right version and reinstall

[ornate elk]

you can overwrite it

but oyu need to delete the venv folder afterwards

before launching the webui-user.bat again

[sinful needle]

okie

[ornate elk]

the other two units are enabled but havent a model selectet:

[sinful needle]

alrighty ima try again

[spiral remnant]

Twitch... Ughhhhhhh

[sinful needle]

whats goin on there

[spiral remnant]

That survived multiple reinstalls

And you were right, the other two controlnet units were enabled 😐

[ornate elk]

did you loaded an image in, to auto adjust the settings? that can enable controlnets too

[spiral remnant]

Yeah let me restart it and see if it does the same when I drop an image in

Yeah that was exactly it, soon as I drop an image in, all 3 units light up

[ornate elk]

ah well 🥲

[spiral remnant]

I'm not fancy, I've just dropped the config to 1 unit so I don't trip and hit my own face again 😛

Thanks CS1o

[ornate elk]

no problem ^^

[craggy granite]

I had to remove and reinstalled stable diffusion a few times to fix this but no chance. whenever I install CUDA, I see this error;

Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check

I know there are ways to bypass this, but then stable diffusion will start using the cpu instead of the gpu. Even if I uninstall CUDA, this continues to appear and only way to fix it is to do deleting all the files including git and pyhton and reinstall them. Any suggestions?

[ornate elk]

hey, whats your GPU?

[wet flame]

yeah check your gpu and cuda to avoid all the problems i faced

check what driver version

someone here recommended me that use 11 after i did that it started creating all problems

then install the compatible torch

If nvidia check this table
https://developer.nvidia.com/cuda-gpus

[craggy granite]

rtx 4070

[wet flame]

them m sure its driver 12.x

and you need to use compatbile torch with it

[craggy granite]

@wet flame can you check dm if thats ok?

[wet flame]

ok

[ornate elk]

you dont need to install cuda for SD

just make sure your on the latest nvidia driver

then follow my install guide from here:
#🤝｜tech-support message

[sinful needle]

Still

[ornate elk]

delete the default model (v1-5-pruned-emaonly.safetensors)

and relaunch

[proper pike]

What am I doing wrong here?

It works on one clip, but will give that error on other clips

[sinful needle]

same

[ornate elk]

put an other model into the model/stable-diffusion folder and delete the default model again

and this is really strange

[sinful needle]

oh i just renamed the repo. been working that way prior to my factory reset

i could change it back, but i doubt itd do anything

[ornate elk]

it will break more

by renaming sd stuff xD

[sinful needle]

whats the original name then ill change it back

[ornate elk]

stable-diffusion-webui

[sinful needle]

alright fine. but ill see if alt models fix first then ill rename

seems to be loading somewhat fine now. waiting for the rest

Unless thats all?:

To create a public link, set share=True in launch().
Startup time: 23.7s (prepare environment: 6.5s, import torch: 7.0s, import gradio: 2.6s, setup paths: 2.9s, initialize shared: 0.5s, other imports: 1.8s, list SD models: 0.3s, load scripts: 1.1s, create ui: 0.8s, gradio launch: 0.5s).
0bbe3f1aa306d4ce40bbfd467b10988e4696b3d4f9c7610b4564517cacd31839
Loading weights [0bbe3f1aa3] from D:\Ai\Automatic1111\Sourcefiles\models\Stable-diffusion\1.5_Sca_Pho__LandscapeBING_--Hyper.safetensors
Creating model from config: D:\Ai\Automatic1111\Sourcefiles\configs\v1-inference.yaml
D:\Ai\Automatic1111\Sourcefiles\venv\lib\site-packages\huggingface_hub\file_download.py:1132: FutureWarning: resume_download is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use force_download=True.
warnings.warn(
vocab.json: 100%|███████████████████████████████████████████████████████████████████| 961k/961k [00:00<00:00, 2.80MB/s]
merges.txt: 100%|███████████████████████████████████████████████████████████████████| 525k/525k [00:00<00:00, 3.90MB/s]
special_tokens_map.json: 100%|████████████████████████████████████████████████████████████████| 389/389 [00:00<?, ?B/s]
tokenizer_config.json: 100%|██████████████████████████████████████████████████████████████████| 905/905 [00:00<?, ?B/s]
config.json: 100%|████████████████████████████████████████████████████████████████████████| 4.52k/4.52k [00:00<?, ?B/s]
Applying attention optimization: Doggettx... done.
Model loaded in 11.4s (calculate hash: 6.6s, create model: 2.5s, apply weights to model: 1.8s, calculate empty prompt: 0.2s).

its no longer automatically booting up the webui page

[ornate elk]

looks good

[sinful needle]

Well it generates things, but why is it no longer automatically booting up the webui in browser?

[ornate elk]

then add --autolaunch also add --xformers

[sinful needle]

weird never had to add that before

besides the xformers

[ornate elk]

it happens sometimes

[mortal notch]

Hello, would anybody be able to help me install stability diffusion on amd? I tried about twice with the youtube guides but cant seem to get it to work. Already uninstalled everything

[ornate elk]

hey, sure, whats your gpu?

[mortal notch]

let me check.

[ornate elk]

i made 2 install guides for AMD so your at the right spot here

[mortal notch]

6800 xt @ornate elk

[ornate elk]

okay, then you need to follow this guide:
#🤝｜tech-support message

[mortal notch]

thats great to hear! i was starting to get a little fustrated lol

I think i followed a similar guide already. It says no module

[ornate elk]

no module named pip?

[mortal notch]

let me reinstall everything and follow your guide. Ill let you know the problems i run into

[willow kettle]

fwiw I thought that guide is great. 6750 xt over here

[mortal notch]

This is the link to download stability fusion right? Just want to make sure incase ive been using the wrong one lol https://lykos.ai/

[ornate elk]

nope

my guide covers all you need to do

Stability Matrix is an Installer Tool which is good for nvidia users, but wont help with AMD (it only can install the slow directml stuff)

[mortal notch]

ok. Ill follow the whole thing step by step then

[ornate elk]

feel free to ask if you have any questions with the steps

[mortal notch]

do i need to redownload git and python?

[ornate elk]

yep

python 3.10.11 64bit

and git 64bit

[fair oxide]

How to import model/lora?

[ornate elk]

download the file and put models into the models/stable-diffusion folder and loras into models/lora

[fair oxide]

When trying to regenerate, I don't have to use point/credits like in other sites, right?

[ornate elk]

if you installed auto1111 localy, so nope

if you talk about the discord bot or other webservices then idk

[fair oxide]

Alright thanks

[mortal notch]

@ornate elk Where do i find %HIP_PATH%bin to add to path?

[ornate elk]

just add %HIP_PATH%bin

it will find the path on its own then

[mortal notch]

sorry, how do i do that? im not very good with computers

[ornate elk]

dont add it to the user variables

[ornate elk]

click here on the Path

the blue

then click on edit

and New

[mortal notch]

ahh i see it

ty

wait, omg i think its working

it says installing requirements for a while now. Do i just wait?

[ornate elk]

yep

it will install everything and open in your browser afterwards

then you have to prompt for something (for example Cat) and just click Generate.
then it will take a very long time doing stuff in the background (can take 15-40 mins)
let it load and you will get the first Image

and after that your ready and can generate images very fast 🙂

[mortal notch]

oh wow, tysm man

[ornate elk]

no problem 🙂

[unreal knot]

Does anyone use sd-webui-discord-ex or another bot like it? I am trying to get a bot up and running but it feels like every time I try one something doesnt work.

[visual pilot]

Hello do you know if is possible to train a pony sdxl lora in onetrainer?

[mortal notch]

@ornate elk its been loading for quite a while. Is it normal for it to take 30 mins+?

[ornate elk]

yep

its nearly done ^^ i can already see the step counter

[mortal notch]

nice. ty again

[ornate elk]

np, thats why in my guide i said 15-40 minutes

[jovial ermine]

Nvidia A40-12Q

trying to get SD to run on our VDI

[foggy basin]

Hi guys, how can I fix?

[fair oxide]

I already put lora here but it didn't show up in the stable diffusion and how do I use it?

[cursive cobalt]

Did you press the little reload icon, in the Lora tab?

The one on the right

[cursive cobalt]

Looks like the lora s gone..? Try to install it?
As a noob within AI, that is what that error looks like

[fair oxide]

I did but it still not there

[cursive cobalt]

Toss me a link to the main one. Lemme check something 🙂

[fair oxide]

Link to what? Main one?

[cursive cobalt]

Ye. The Lora

[fair oxide]

Oh, okay

https://civitai.com/models/22298/horikita-suzune-classroom-of-the-elite

https://civitai.com/models/196472?modelVersionId=220900

[cursive cobalt]

Obviously didn't fine tune the actual promts, but make sure its installed in
D:\A1111\stable-diffusion-webui\models\Lora.
If they are, reinstall them.
Worked fine for me

@fair oxide

[fair oxide]

Damn... I already put it there...

D:\AI\stable-diffusion-webui\models\Lora

[cursive cobalt]

Try to reinstall them

[fair oxide]

I just have to cut and paste it right?

[cursive cobalt]

Not sure. Just paste them in another temp folder for now

then get em back in

[fair oxide]

Oh okay

Still won't work...

[cursive cobalt]

That is very weird. If that didn't work, its beyond my knowledge. Sorry 😐

[fair oxide]

It's okay, thanks for helping

[cursive cobalt]

https://tenor.com/view/attempt-there-was-an-attempt-rainbow-star-space-gif-15526647

[fair oxide]

It's a pity since I want to use it to generate some temporary art for my game

[cursive cobalt]

I mean... It works for me, so if you want, head to #🍥｜anime , and give me a list of your promts, and i can see what it can do. Using the Pony-SDXL merge model atm.

At least until your issue is fixed.

So you at least have something

As long as what you're making, isn't anything breaking any laws 😉

[fair oxide]

Nah, it's only temporary until I have the money to hire artist

But thanks I'll try to consider it

[cursive cobalt]

Oki

[fair oxide]

hi guys I am running into a problem
need help
trying to train a lora
and it gives me this: PermissionError: [Errno 13] Permission denied: '/config_dreambooth-20240521-111115.toml

[ornate elk]

Make sure You have a compatible model selected

In that case it has to be 1.5 based

[craggy latch]

is it normal that i have 2 progress bars while generating 1 picture?

[ornate elk]

Yes

One is for the steps and the second is for the total progress

Like when using hires fix or other stuff

[craggy latch]

thanks

[vestal anchor]

I tried using scaling but encountered an error saying "not enough memory" Scaling required 6 GB, but only 4 GB was available, even though my graphics card has 8 GB. I don't understand why this is happening. Do I need to unlock more VRAM on my card, or is there a way to increase VRAM access for sd?

[ornate elk]

hey, whats your gpu, and whats inside your webui-user.bat?

[jovial ermine]

Hi again, I tried installing Comfyui instead of A1111. I am getting this error. Any thoughts?

[idle bolt]

I get an error message that a url: /api/models/stabilityai/stable-diffusion-xl-base-1.0 is missing when i start lora training local on windows without internet.

I always get this message when I don't have internet on. I use Kohya locally under Windows. can I download this url: /api/models/stabilityai/stable-diffusion-xl-base-1.0 file manually? If so, which folder do I have to put it in? So Kohya always needs internet. I thought it would work completely offline? It works with internet because he is still downloading something

Thanks.

requests.exceptions.ConnectionError: (MaxRetryError("HTTPSConnectionPool(host='huggingface.co', port=443): Max retries exceeded with url: /api/models/stabilityai/stable-diffusion-xl-base-1.0 (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x0000026DE9F71630>: Failed to establish a new connection: [Errno 11001] getaddrinfo failed'))"), '(Request ID: 6ead4e04-602f-4b37-b662-a6b1f0a3f64d

Can I download the file myself? And in which folder do I have to put them?

Kohya_ss

[marsh plume]

Cannot load my saved deforum settings, how to fix this ?

[vestal anchor]

Gigabyte GeForce RTX 3050 EAGLE OC 8GB GDDR6 128bit DLSS

file "C:\Users\korde\Desktop\stable-diffusion\stable-diffusion-webui\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "C:\Users\korde\Desktop\stable-diffusion\stable-diffusion-webui\modules\sd_hijack_optimizations.py", line 260, in split_cross_attention_forward
raise RuntimeError(f'Not enough memory, use lower resolution (max approx. {max_res}x{max_res}). '
RuntimeError: Not enough memory, use lower resolution (max approx. 1600x1600). Need: 6.0GB free, Have:4.2GB free

[royal helm]

Is this install instructions for a full SD install or is this just Zluda? Completely new dude just looking for insstall instructions bc I haven't been able to get it working yet.
amd RX6700XT ryzen 5 3600, Nobara linux (fedora)

[ornate elk]

you need to edit the webui-user.bat
at the line COMMANDLINE_ARGs= you have to add: --xformers --medvram-sdxl --no-half-vae
then save and relaunch the webui-user.bat

that should give you a much better performance

[ornate elk]

hey, this is for a full SD webui install + Zluda on Windows

[royal helm]

The rest of the community pointed me here. Maybe someone can point me in the right direction

[ornate elk]

im not a linux user so i cant help much with it sry :/
but if you have trouble getting rocm to work on linux then checkout this guide:
https://github.com/xzuyn/ROCm-Guides

[vestal anchor]

it worked thanks CS1o

[royal helm]

Cheers anyway. It's gonna be a headache and I'll probably be back

[foggy basin]

hey guys, I’m almost getting ComfyUi to work , can anyone kindly guide me on the missing steps ?

[vestal anchor]

OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 0 has a total capacty of 8.00 GiB of which 0 bytes is free. Of the allocated memory 7.21 GiB is allocated by PyTorch, and 1.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

I tried to resize from 1024x1024 to 2048x2048 and I got this error. Does anyone know what I should do based on this error?

[ornate elk]

How do you try to upscale it?

Thats a heavy jump in resolution

Your GPU don't have enough Vram to do that

[vestal anchor]

I just uploaded a picture to img2img and tried to resize by 2
Is there a better way to achieve higher resolution?

[open fox]

Maybe a tiled upscale. Then you create the image by upscale n-parts with each a lower resolution. But i think that supir, ultra upscale etc. should already use this technique to avoid out of memory

[ornate elk]

You need to use the SD upscale script at the bottom of img2img

Set the denois below 0.3

[waxen pebble]

ive had success training LORAs in Kohya ss.. but have recently been testing out my ability to train full checkpoints with Dreambooth... to no avail.

research into whats going wrong is also leading me to question... is there an advantage to training the full checkpoints or is LORA the way forward?

I was thinking that by training some style concepts into a dreambooth i would have better luck combining that model with addional loras at generation stage, where combining multiple loras with signficiant influence can be tricky....

anyone have input as to lora vs dreambooth training full models??

[prime sand]

hey im trying to use stable diffusion yesterday it works well but today i have this error

[bright light]

I don't know if any A1111 devs check this discord but, since the recent A1111 updates, this plugin doesn't work anymore, I don't know if the protocol that works with the api has changed recently, but Stable.Art doesn't work since 😦
https://github.com/isekaidev/stable.art/issues/78

[acoustic lynx]

Hi ! I'm making a comfyui workflow and in the workflow I need to get CLIP Score between image and the prompt. Does somebody have already done this within comfyui or is there any custom node available for it as I didn't find any.

[vestal anchor]

Thanks that really helped

[meager oyster]

Hello.

I got that error message.
The installation is fresh. I have never encountered such a thing before.

It's Automatic's Stable Diffusion Web UI.

Error is produced after double clicking the "webui-user.bat" file.

[ornate elk]

hey, can you show a full cmd log?

also whats your gpu and whats inside your webui-user.bat?

[wind thunder]

can someone help^me ?

what do i have to do

maybe its because i changed the default web browser

@ornate elk ? pls

[ornate elk]

install python 3.10.11 64bit and then delete the venv folder and relaunch the webui-user.bat

[proper pike]

Does anyone know why its so jarring to look at?

[ember mortar]

im having a problem with adetailer, it keep doing the face with a very low detail

[wind thunder]

?

[ornate elk]

when did you installed the webui ?

[wind thunder]

like 2week ago

[ornate elk]

open up a cmd and run:
pip cache purge

also do you have any third party antivirus installed?

[wind thunder]

What do you mean by third party

i have like the malwayre antivirus and the windows defender i thinl

[ornate elk]

malware anitivirus?

[wind thunder]

i mean this sry

[ornate elk]

ahh okay

you have to disable it when installing python, it could block something

[shell crystal]

Pretty new to stable diffusion, so I want to ask how do you go about generating an image with multiple loras? I use controlnet/open pose and although I get two characters, they fuse the same characteristics on both characters

Would like to know if there's a good guide or some resource I could take a look at for someone new like me

[wind thunder]

maybe i could disable the windows defender no ?

[ornate elk]

it doesnt block python

so leave it enabled

[wind thunder]

strange

[ornate elk]

hey, that can be done with the multidiffusion extension or the regional prompt control extension

[wind thunder]

so whats the solution do i have to reinstall it completly or just search the potential element that block the module installation

[ember mortar]

im having this problem with adetailer and the face it make is awful idk why

[ornate elk]

You need to reinstall python mostly

But you can also try a reinstall of SD via my guide

[sinful needle]

webui is still not automatically opening in browser despite the args

[ornate elk]

make sure your wbeui and the extensions are updated

[ember mortar]

updated both but i still have the same problem

[ornate elk]

Which model?

[ember mortar]

face_yolov8n

idk if im adding the right prompt in adetailer, im still new with this

im trying with 2 chars

yesterday it was normal but today idk why it always happen

[sinful needle]

set COMMANDLINE_ARGS= --autolaunch --xformers --no-half-vae

[ornate elk]

Try without setting anything in adetailer

[ember mortar]

ok

[sinful needle]

there was something at the bottom but i forgor

[ember mortar]

does a lora affect adetailer?

done without prompts in adetailer and still bad

[ornate elk]

What model do you use and sampler

And I don't mean adetailer model

[ember mortar]

checkpoint and lora?

[ornate elk]

Don't add the lora to adetailer prompt

[ember mortar]

is fun cause i added to one and it worked but not for the other one

the checkpoint is autismmixSDXL_autismmixConfetti

[ornate elk]

Ah pony models

[ember mortar]

yep

also again

i added the lora in adetailer and in one it work but not in the other

i think is a problem with the lora

still i tried without lora on that one that don't work and still do the same

[ornate elk]

Lora's need to be compatible with pony

Not every sdxl Lora works on pony

And not every pony lora works on sdxl

[ember mortar]

both are compatible without adetailer, at least for what i get are good

so, is it possible to fix it? ._.

[ornate elk]

Do you have the lora in the prompt or Adetailer prompt?

[ember mortar]

first, it was at the prompt, now i tried in both and at least in one work

and could be fixed adding boht lora in both prompts, still it sometimes ignore some details

[ornate elk]

Hmm okay

[ember mortar]

but at least it looks good

the other one i can't fix it

it keep doing the same

btw im using regional prompter too

but im pretty sure the breaks are okay

[jagged glade]

how to fix this quickly? ( i need video tutorial pls)

[cursive cobalt]

Beep boop and morning. @ornate elk Do u know, if one can have several folders, spread out on different drives, that 1 WebUI can call upon for models? Like D:\Folder1, E:\Folder2, etc

[turbid echo]

can I upload my image and make a video out of it?

[cursive cobalt]

Oh, and another question. Can i change the text of these, without altering their functions?

[ember mortar]

is there a prompt for always get 5 fingers for hands or just try?

[cursive cobalt]

You can try the Perfect Hands lora, or use ADetailer and try to fix in img2img

[ember mortar]

okey ty

[undone robin]

sdxl api
error code: 503
upstream connect error or disconnect/reset before headers. reset reason: connection failure

[ornate elk]

Deleting the dreambooth extension and then the venv folder will fix it

[jagged glade]

Ok thnx

[ornate elk]

You can just use them normaly. And control them over their strength

[ornate elk]

NP, you should also use xformers for better performance

[jagged glade]

I will ask someone about xformers I don't know anything about it

[ornate elk]

What's Your GPU?

[undone robin]

500 {"id":"bd539ed1820641242dae05ae9114244a","message":"Unable to find corresponding account in internal DB. Contact [e-mail] if this issue persists.","name":"server_error

[jagged glade]

3060 12gb

[ornate elk]

Okay then you just have to edit the webui-user.bat, at the line commandline_args=
You add: --xformers
Then save and relaunch it

[jagged glade]

👍

I am still learning about these things

[amber hamlet]

hi everyone

how can i fix this issue

[ornate elk]

No problem. If you want to train stuff, don't use dreambooth. Better use the Kohya_ss tool. Its a separate training tool that won't break your Stable diffusion webui

[ornate elk]

Hey, I see you used the zip variant for installing SD.
Thats not the recommended way.
Better follow my install guide from here and see if it works then:
#🤝｜tech-support message

[amber hamlet]

okay thnx let me try that

[ornate elk]

Yes thats possible by editing the webui-user.bat at the line commandline_args=
There you can add: --ckpt-dir "D:/Path/to/ModelsExample"

[cursive cobalt]

Is it limited to only 1 extra folder, or can i add as many as i want, theoretically?

[ornate elk]

I didn't tried multiple, but if that won't work then you can always use Symbolic Links as many as you want

[cursive cobalt]

Cause i want to retain the root folder, and those that are in. But I'm running out of space on this drive

And to avoid reinstalling it........ again

I wanna add extra folders across the drives

Symbolic links, in Auto1111..? Wat?

[ornate elk]

In the folder

Not in auto

[cursive cobalt]

https://tenor.com/view/i'm-lost-evelyn-lozada-basketball-wives-i'm-confused-i'm-at-a-loss-gif-9885762462977519313

I am not a smart cookie, sorry 😦

[ornate elk]

NP, first try it with the --ckpt-dir

[cursive cobalt]

Right. So

--ckpt-dir D:/Path/to/ModelsFolder1
--ckpt-dir E:/Path/to/ModelsFolder2

?

Equalling this, ye?

[oblique parrot]

I'm getting errors from the API:
$VAR1 = {
'name' => 'internal_error',
'id' => '765736180dab3a9a2ad823cb226401ea',
'errors' => [
'An unexpected error occurred while trying to reach an upstream service. Please try again. If the problem persists please contact support and include the id of this response.'
]
};
Request failed: 500 Internal Server Error

[quartz creek]

Is the API down?

on Platform.stabiity, also getting db errors on our account

[oblique parrot]

Also of note: when I look at my account page, everything except my name is grey bubbles.

[quartz creek]

Yep same

[oblique parrot]

wellp, I put in a bug report for one of the failures.

I'm not gonna do all of them. I have eight bugs open being ignored already, and I would have to submit like 120 of them if I did every one. (I made a perl script that submits a creative upscale request and waits on the results, and if it gets either a 403 or the output is blurred--blurriness detected with imagemagick--it tries again, because the stuff I'm asking for is all work-safe. This started happening after the 3,947th attempt to upscale a picture of a female vampire on a cathedral rooftop--we're not talking Vampirella here, we're talking Victorian dress up to the neck and so little skin showing it wouldn't even bother a pilgrim. They're so sexist.)

And don't anyone worry, my script sleeps for 1 second between every try. It doesn't clobber them fast. Not that it could because the tries themselves take about a second or two.

This is weird. Magnific is also just spinning. I didn't think Magnific used SD but with the way it's hanging right now, I think I was wrong.

[cursive cobalt]

@ornate elk Right'o, sorry about the afk Door rang. Running the webui bat file now, to see if it loaded.

Mass error flow

[ornate elk]

You didn't used "" for the paths

[cursive cobalt]

Oh, it needs "". Oh, brb then

[ornate elk]

"Path"

[cursive cobalt]

That seemed to work. Fanks :3

(I named the E Drive file manually to verify amongst the mess of models i has)

[ornate elk]

Ah nice

[amber hamlet]

😦 what did i do wrong mugiwara senpai

[ornate elk]

Oh, python wasnt installed correctly.
If you have an third party antivirus Programm running, that breaks the python install

You have to uninstall python. Then turn off the antivirus Programm and then install python again

[zenith kite]

No cascade in stable models discord? I'm trying to set up cascade in A111 forge, native so I can use it to upscale. Has anybody had any luck with that? My workflow is tied to a111 so I don't really want to migrate to comfy if I can help it

[main whale]

Is the stable diffusion 3 API down?

[cursive cobalt]

I forgor. I had another question. is it possible to change the name of like these 2, but not change their function? Its not that i can't remember what height does, its more CFG, i wanna change to something MY brain can easilier understand at first glance

[ornate elk]

Stable cascade for upscale?

Afaik its not supported in Auto1111. But there is this thing:
https://github.com/umisetokikaze/StableCascade-webui

[ornate elk]

Maybe in the ui-config.json

But I haven't tested it

[cursive cobalt]

Apparently you can edit it in a localization file... Now, the issue is. The ones i found are all but english, so i have to edit them all xD

So this is gonna take some time .

[ornate elk]

Use notepad++ it has a search and replace feature

[cursive cobalt]

Well yeah, but I'm not german, and i don't understand german, so i gotta replace all the german back to english 😄

[ornate elk]

You don't need to change the German stuff

It won't be used if your webui is english

In your case I would just write me a txt note with an explanation what CFG does, instead if changing so much stuff

[cursive cobalt]

But shouldn't the ui config file look something like the german one up there?

Cause that looks dangerous to edit

[cursive cobalt]

Speshal Brian requires speshal text :3

[ornate elk]

Yea you can't edit the stuff in front of the values

It would break stuff

[cursive cobalt]

Yeah. thats what i thought

So i figured a localization edit would be better

and more safe

[meager oyster]

Python is 3.10.6, launch argument: - - xformers ; in the bat file haven‘t put anything else: python directory, git, venv_dir are empty; last line in bat calls for webui.bat

I did not change the venv directory cause it sets one in the webui.bat

[ornate elk]

Okay, can you still show the whole cmd log when launching the webui-user.bat?

That would help me find the error

Also recheck my guide if you did everything correctly:
#🤝｜tech-support message

I'm AFK (20mins)

[zenith kite]

Ok, thanks for that. If possible I'll gen images with cascade and do post in forge. I couldn't seem to find any cascade stuff on github with img2img

But I like the grimy 512x512 images from sd15, they're perfect for playing around with for huge upscales

[cursive cobalt]

Welp this is annoying... I've added the localization file into the correct folder, but the ... client..? Isn't finding it

Found out why. I saved it as a .txt file, but needs to be saved as a .json file

[amber hamlet]

@ornate elk thank you for solving my problem !!! you're gonna be king of pirates for sure

[meager oyster]

After that I moved the folder to C: and added the other two arguments (- - no - half - vae and - - medvram - sdxl; I am aware that all the - need to be without spaces, writing that because of the iPad doing weird things with them)

Still the same error (the paths have changed *ofc)

[jagged glade]

i never used dreambooth just added it for later projects 6 months ago

[ornate elk]

Ah well, then check if your webui is updated too

Maybe your on an old version

[cursive cobalt]

\o/ Yay. It worked

[ornate elk]

Delete the venv folder and relaunch the webui-user.bat

[meager oyster]

That will download everything again.

Is that intended?

Oke.

That will take some time though. C U when it has finished

[ornate elk]

It shouldn't take long as the files got cached at the first download

[meager oyster]

Oh, did they?

Let‘s see.

Lol. Connection currently being broken since my local internet connection is disrupted at the moment. Will try to fix it the next few hours.

[ornate elk]

Okay, then delete the venv folder again at the next try

[meager oyster]

Yep, will do.

[ornate elk]

If your using a Telekom Router, unplug / replug can help ^^ they broke my connection two days ago with a firmware update over night

[meager oyster]

The router is from Asus (one of the fancier ones my brother bought it two years ago)

And it is Telekom but something is under construction so I cannot influence it in any way but to call them.

I mean the internet is provided by the German Telekom.

[cursive cobalt]

Hey Guyses. Did you know, if u want to make your own localization, the WebUI has a function to create your own template you can edit in? Just remember to save as a .json 😄

[ornate elk]

Ah the Asus Routers are nice, much better features then the Telekom Speedports

[meager oyster]

I do not know. He *said he bought it for the ping (?)

[finite token]

is stability down? Logged into my account and cant see any credits or previous payment details - I only renewed last week so know I havent used all my credits up

[loud pollen]

It is down

[finite token]

Thanks - got no info in my account so didnt know if it was just me. Any ETA from them, havent seen anything?

[ornate elk]

A real gamer ^^

[meager oyster]

But then you wouldn‘t use WiFi, would you?

[ornate elk]

No way xD

Ethernet is the only answer for the best Gaming experience

[cursive cobalt]

New discovery. The "--ckpt-dir "E:\Stable-Diffusion\Models" " only works for models it seems, adding a 3rd line of "it" breaks it

Tried adding the path for Lora's.

[ornate elk]

For loras its --lora-dir

[cursive cobalt]

Oh

[bright light]

(I know it's not polite, but I'm reposting this here, just in case...)

I don't know if any A1111 devs check this discord but, since the recent A1111 updates, this plugin doesn't work anymore, I don't know if the protocol that works with the api has changed recently, but Stable.Art doesn't work since 😦
https://github.com/isekaidev/stable.art/issues/78

[ornate elk]

There are also:
--embeddings-dir
--controlnet-dir
--esrgan-dir

[meager oyster]

For me stability and reliability are the only reasons why my computer is connected via ethernet. What he chooses is on him.

[cursive cobalt]

?

Oh right. Cause dumb brain didn't connect "ckpt" = ChecKPoinT

Told u brain doesn't quite work right xD

[ornate elk]

No A1111 dev checks this Discord. They have their own.
Best is to make a github issue in the extensions site or the Auto1111 site depending what broke it

[bright light]

Hmm interesting, thank you 🙂

[cursive cobalt]

You know what else isn't polite? Eating a metric load of beets, using the toilet and not flushing, making your family think you're dying.

[ornate elk]

No problem 🙂 and thanks for the link. Didn't knew that extension

Would be nice for Krita. But it only works with comfyui as backend

[bright light]

I have an old version of Forge that's 2 months old, it still works with it, hence the fact that I think they've updated the way the API works, that a lot of plugins have followed too, but this one hasn't been updated for a long time, but it's still by far the best in its field...

[ornate elk]

Oh I see, updated last year

Then you would need to git clone an older version of auto1111 to get it working

[bright light]

I know, but I really don't want to have 50 versions of A1111, to systematically close/restart after each gen, then inpainting.... It's a shame, but I'm sure it can't be much, and that the A1111 devs can fix it quickly :/

do you have link for official A1111 discord? there are lots of them on the net, I'm not sure which is the right one.

[ornate elk]

nope, as the discord is only for Devs, they wont deal with issues there. Make an issue on the github repo and they will look at it. also make an issue on the extension repo, maybe the dev wakes up xD

[bright light]

I was hesitating. I wonder if it wouldn't be much faster if I tried to update this plugin, I know nothing about Vue and .Js. But it can't be worse than C# .net

[ornate elk]

yea, why not fork the project, let chat gpt look over it and make a workaround ^^

[bright light]

I'd thought about it, but the problem is that this plugin actually works, it's the functioning of the A1111 API protocol that remains unknown, for me as for LLMs.

https://tenor.com/view/falafel-gifalafel-hazfalafel-icanhasfalafel-reaction-gif-reaction-פלאפלאנילקבליכולפלאפל-פלאפל-אנילקבליכולפלאפל-digging-gif-13411844

[ornate elk]

hmm, then maybe checkout code of other extensions how they use the updated api protocols

[bright light]

Yeah, exactly. That's why, out of laziness, I was sadly looking to see if any devs were listening on the A1111 side. Obviously without much hope

But I'm in a digging mood x), so I'll try to do this on my own. If it helps others, by the way

[ornate elk]

Nice, let us know how it goes 🙂

[meager oyster]

It works now. Thank you so so much.

[ornate elk]

Perfect 🙂 no problem

[slender hollow]

Sorry, I'm totally new here. One question: how can dreamstudio start? when I want to log in, he always tells me to verify my email address.

[bright light]

@ornate elk
I have the beginnings of a solution 🙂
https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/15865

[ornate elk]

Awesome!

[upper linden]

need help getting zula to work, it suddenly gives me this error: Launching Web UI with arguments: --use-zluda --medvram-sdxl --api --update-check --skip-ort --ckpt-dir F:\webui\Ai\stable-diffusion-webui-directml\models\Stable-diffusion --lora-dir F:\webui\Ai\stable-diffusion-webui-directml\models\Lora

rocBLAS error: Cannot read C:\Program Files\AMD\ROCm\5.7\bin/rocblas/library/TensileLibrary.dat: No such file or directory for GPU arch : gfx1036

rocBLAS error: Could not initialize Tensile host:
regex_error(error_backref): The expression contained an invalid back reference.

[upper linden]

i also tried to reinstall AMD HIP SDK 5.7 to see if i can get the file but it didn't work

[ornate elk]

Disable the Igpu in device manager

Radeon TM Graphics

[upper linden]

ty

[steel geyser]

trying to use gif2gif + reActor but the output is just a jumbled flashing mess, any idea what im doing wrong? i just want shrek to be in django unchained 😦

the output ends up looking nothing like the source gif at all

[steel geyser]

Have you tried reinstalling a compatible version of python and git?

I had a similar issue but doing that fixed it for me

[verbal sundial]

Anyone else getting issues with image browser like this? or is it just me?
idk if it makes any difference but my stable-diffusion-webui folder that has everything in it is in a separate hard drive

I can browse images but can't see the prompts I used... maybe something to do with metadata not being read properly??

[steel geyser]

I personally use it to pull various models/extension/variants of SD as it makes the process much easier for me

Seems like alot more questions than answers on here haha

But I guess that's life

[ornate elk]

normaly not ^^

but the latest questions are to specific stuff

would need to test the extensions to help

[green merlin]

hello , can someone please save me ? i have literally followed every youtube video , asked gpt , read comments but i keep getting these 2 errors the first on the normal webui , the second on the v1.0.0-pre that auto install everything .
i did add the command to skip torch and didnt work . i have tried everything with pip but nothing . i have been trying for 12 hours now its 00:09 as i speak and i am feeling desperate.
if you know any solution please i would be very thankful .

[ornate elk]

hey, whats your GPU?

[green merlin]

a 4060 rtx

[viral knot]

CS1o, the true MVP of this channel ❤️

[ornate elk]

okay, then delete the whole stuff.
and follow my install guide.
before you start open up a cmd and run:
pip cache purge
then make sure any third party antivirus programm is disabled and that your GPU driver is updated.
also make sure you have a stable Ethernet connection.
Here is my install guide:
#🤝｜tech-support message

[green merlin]

ok thanks i will git it a try

[sinful needle]

so the prompt is:

city at (night:1.4), ground level shot. Fog rolls in from the side streets, red lighting from the windows, (moving clouds:1.1).
. Hyper realistic photo, hirez, HD, 8K, hyperdetailed, hyper quality, artstation winning award, bright highlights, deep shadows, perfect composition, soft shadows, Fantastic Scenery

but i keep getting far away views. and no, the new lines dont really make a difference i tried it.

[clever needle]

there is this loading sign in my ui and i cant genrate anything

im using froge ui

[clever needle]

every thing loading and have a timer next to it

[sinful needle]

more animatediff errors

[hollow kayak]

Stable diffusion is out of the blue refusing to work at all. I was hoping someone could say if this log could shed any light on the issue. Thank you.

[solemn moon]

DID YOU UPDATE DRIVER

[chilly sandal]

(pls correct me if I posted in the wrong channel)
why is Regional prompter doing this to meee?
I separated in 3 regions, and it messes up either the last region or the last 2 at once.. any clues why this is happening?

tried changing the sampling method, model, hires, nothing works...

[hollow kayak]

Yes, i updated my video driver.

[wary bobcat]

Hi there, good day. I run stable diffusion from comfyUI, when I process a 1 sec video 720x1080 25fps, it takes more than 30 min to do so. I use nvidia RTX A6000 GPU with 48GB vRAM. and the comfyUI has only 2 controlnet, 1 animatediff model, and one base model. any one has a deal how can I optimize the performance? thank you so much. my configuration and running status is attached.

[hallow latch]

Hey, I need some quick help.

I just re-installed my PC and installed git and stable but when I hit the webui-run bat file .. nothing happens. I see the cmd windows for 0,3 seconds or something but nothing happens.

Do anyone know the reason why?

[ornate elk]

You need python 3.10.11 64bit too

[ornate elk]

Try an other browser. Also make sure the webui is whitelisted in any browser adblocker

[hallow latch]

I think I got it to work somehow, I installed stable inside git.. but did cd %userprofile% instead and was working after that. Not sure why and I did not get any option to select "Get Path .. " for python

[ornate elk]

You tried used a Lora that is maybe not a lora

[hallow latch]

I have not updated stable in a long time.. What is "Schedule Type" ?

[ornate elk]

Before 1.9.3 we had samplers like
DPM++ 2m Karras or DPM++ SDE and SDE Karas.
These got split now into sampler and scheduler.
So Karas is now separated

If you leave it on automatic it will always use the best for the sampler

[hallow latch]

Okay cool, I am not sure if my CPU does any changes to the speed. I went from a AMD Ryzen 7 2700x to and Amd Ryzen 7 7800x3d now 🙂

[ornate elk]

A beast of CPU ^^ I have the same

[hallow latch]

Yeah so far Im super happy. It's more for gaming as my old CPU was very bottlenecked. I played CS2 just few hours ago and went from average FPS 140~ on 1280 res to 300+ fps with 2560x1440 on highest settings 🙂

So far I am super happy for this upgrade and I can only hope that my GPU will survive for couple years 🙂

[ornate elk]

What's your GPU?

[hallow latch]

ASUS 2080ti 11 GB

[ornate elk]

Ah nice

[hallow latch]

It is 5 years old, so I hope it will survive some years before I need to buy 🙂

[ornate elk]

GPUs don't fail that often

If you dust them off from time to time they keep running and running

[hallow latch]

I think it's kinda "random" sometimes you can be lucky and all .. But I just need to get help from asus now so that I can replace the termal padding and etc 🙂

What GPU do you run with you ryzen?

[ornate elk]

Oh Asus and support will be tricky, if you watched the GamersNexus Video a few days ago xD

[ornate elk]

I'm running a AMD Radeon 7900XTX

[hallow latch]

I have not seen but seen some tiktok videos etc about something. So far I am getting okay help, they just need the serial number so far.

[hallow latch]

How is that going for you Gaming wise and heat wise?

[ornate elk]

Gaming wise its awesome, can run every Game maxed out on WQHD, and even stream Games at 4k 60 to the TV for sp couch gaming.

Heat isnt a problem at all. The GPU only gets around 75celsius when playing Cyberpunk 2077 on Max settings.
In most other Games its like 60-65c

[hallow latch]

That is nice ^^ Next GPU may be an AMD in the future but I don't know 🙂 I've had good experience with my 780ti, 1080ti but this 2080ti has been "Meh" and may switch to AMD next time 🙂

I have problem with my 2080ti going up to max "83c" when just doing AI stuff .. like high-res and a large batch so the fans goes crazy xD hahaha

[ornate elk]

Oh wow thats to hot for this task

[ornate elk]

I was on a GTX 1080 before I bought my new PC last year in August

Great card

[hallow latch]

Yeah 83 is the max it gets and it's up there quite a lot sadly :/ So wanna change the termal padding and such on them. But I've read on many reddit pages and other pages that many 2080ti users has this problem sadly

So my experience with this serie has been "Meh"

[ornate elk]

Ah damn :/ make sure you can control your case fan speed.
People without a fan control software or controller forget to set the fans to Auto in BIOS or make a good fan curve there

My new case has 10 fans xD 3 are from the AiO on the ryzen.
Airflow is perfect

[hallow latch]

Yeah, I use Fan Controll on my PC but I am a newbie in the fan controlling and etc..

I know people buy like an extra thing so they can control all fans and such.

I only got 3 in front, 1 CPU fan, 1 in the back atm.

[clever needle]

i love you <3

[hallow latch]

I should invest in some new fans 😄

[ornate elk]

Fan control is a good software

I have a hardware fan controler from the case + aio, same brand, works very good

[hollow kayak]

I seem to have fixed my issue by changing browsers.

[sinful needle]

Another animatediff error.

Only happens when I enter a prompt longer than just a few words. "bouncy ball" will work for instance. I tried an even longer prompt than what error'd and it worked? It's being really picky with the prompt size.

[nova berry]

How to avoid square borders when expanding the image size?

I use controlnet tile + Ultimate SD upscale

[fiery trellis]

OK

[cursive cobalt]

@ornate elk Fren. disaster. Is not finding my Lora's on the 2nd folder

The models are 👍 , the Lora's are not

[hallow latch]

why so much command lines?

[cursive cobalt]

Cause i have seperate folders for Models and Lora's, (thats 4 of them), rest is per advice of CS1o 🙂

[hallow latch]

Why? 🙂

[cursive cobalt]

Cause I don't have enough space on 1 drive...?

[strong hedge]

Has stability released their API server? I'd like to self-host their exact API so I dont have to double-up on logic in development.

[gaunt depot]

Hi Can you please help with API?

I am using Stable Diffusion with aicontentstudio of planifyx but my images and videos are not working.

API key is connected and API key test is Success

what could be the reason? can someone help me to fix this

I've encountered issues while integrating the Stable Confusion API with Planifyx AI Content Studio.

When testing my API key, I received a successful result, indicating that the connection is established. However, I'm facing difficulties with the AI Image and AI Video functionalities within the platform. Despite the successful connection, these features are not functioning as expected.

Could you please provide guidance on what kind of issues could be causing this discrepancy? I'm eager to resolve this matter and ensure seamless integration between Stable Confusion and Planifyx AI Content Studio.

[lusty sage]

Hey does anyone have experience installing and using the BMAB extension? I've been trying to get it to work but with no success. Let me know:)

[ornate elk]

Hey, I installed it and tested it a bit

[lusty sage]

And it worked okay? I watched a video in Korean and it mentioned CUDA and pytorch are needed, which is maybe why it wasn't working for me

[ornate elk]

whats your GPU?

[lusty sage]

1080gtx

[ornate elk]

i used it on AMD witz Zluda

[ornate elk]

that should work easily then

[lusty sage]

That's what I thought too

[ornate elk]

do you get any error?

[lusty sage]

No errors, it just doesn't seem to activate, after processing in the png info there's nothing to suggest it started

Are there more steps to activate it? I figured it'd just be "enable BMAB" and tweak what I want

[visual orbit]

Hey guys, are these results for a used 3090 ok?

[ornate elk]

yea thats it

is your webui updated?

[lusty sage]

Yeah it is

[ornate elk]

have your tried to just change the contrast and brightness?

IClightning wont work for me as it needs more stuff