🤝｜tech-support

[storm sail]

ok, ok, well, again, i apologize if i had offended anyone, i got one more idea here https://github.com/intel/openvino-ai-plugins-gimp/releases i was using this before and wasnt all bad, it rendered with decent speed tho not the best graphics (maybe something to do with model or something), ill try this route once more it looks like this been updated

shot in the dark but least it says point blank "you no longer need to manually install openvino" so i can be sure im getting my boost here . . . had my sight set on stable boy but if idk whether i gots openvino or not for sure . . . or, fingers crossed, i get best of all worlds

only now idk how to compile from source code so . . . and that looks like only option for this thing

anyhow, ill be bck to bitch more later, count on it 😛 til then, yall be well 🙂

[storm sail]

anyone can tell me how i might put together that source code there for them openvino-ai-plugins-gimp 2.99 r2? im on windows btw

i would just download from the green button on that other page related but i dont trust id be getting the very newest there

[karmic crown]

https://github.com/intel/openvino-ai-plugins-gimp

clone with git

git clone https://github.com/intel/openvino-ai-plugins-gimp.git

and run

install.bat

or download from https://github.com/intel/openvino-ai-plugins-gimp/releases/tag/v2.99-R2
unzip and run install.bat

[vast hull]

anyone know how to put stable diffusion on taskbar

[fierce hollow]

give a detailed instruction on how to enlarge the finished photo in resolution? I have an order to increase the quality and resolution of the photo. How do I do this?

[karmic crown]

easy
create a shortcut
https://www.easytechguides.com/pin-a-batch-file-to-the-taskbar-or-start-menu-in-windows-10/

but be sure to add full path to your stabile diffusion folder in start in

for example
c:\ai\stable-diffusion-webui

[vast hull]

i messaged u real quick @karmic crown

i think it deleted my folder?

[karmic crown]

1. create a shortcut like described in the link (let it be on the desktop)
2. right click the shortcut and end find "start in" field and add the full path to folder
3. try to open it
4. if it works add to taskbar as described in the link

[vast hull]

it does the start in but im confused on what exactly to put in target

@karmic crown

[karmic crown]

where your webui-user.bat is?

[vast hull]

C:\Users\HIGHDEEUHH\Documents\StableDiffusionAuto11111

[karmic crown]

than put that in start in

[vast hull]

this is the target url to desktop shortcut webui

C:\Users\HIGHDEEUHH\Desktop\webui-user - Shortcut.lnk

[karmic crown]

its ok

it should be like that

is everything working?

[karmic crown]

target should be path to webui-user.bat

you have 2 shortcuts

you just need one

[vast hull]

can i call u real quick and screen share? kinda confusing via typing

@karmic crown

[karmic crown]

C:\Users\HIGHDEEUHH\Documents\StableDiffusionAuto11111

[vast hull]

put that as target?

[karmic crown]

waith

like this

[vast hull]

cmd /c "C:\Users\HIGHDEEUHH\Documents\StableDiffusionAuto11111" ?

[karmic crown]

cmd /c "C:\Users\HIGHDEEUHH\Documents\StableDiffusionAuto11111\webui-user.bat"

[vast hull]

ok

[karmic crown]

and C:\Users\HIGHDEEUHH\Documents\StableDiffusionAuto11111
in start in

[vast hull]

it works u the man!

how do i put a custom icon on it 😄

[karmic crown]

i guess

i am using windows only when i need 🙂

a few times a year

i forgot many things since i first switched to linux, than to macos

[karmic crown]

@vocal burrow@ornate elk check out this

https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/16113

I accidentally stumbled upon this bug, while i was testing https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/16092

😳

[sinful igloo]

amd gpus still way behind nvidia for image generation?

or are there changes

[ornate elk]

There are changes since Zluda release in February, you can generate pretty fast on windows now

Also WSL2 beta dropped a few days ago

[ornate elk]

Very interesting find!

[spring pike]

having lag spikes mid loading and when its about to be do loading , there for a short time

[solemn atlas]

I'll try it , thank you so much

[steep pawn]

Hello everyone, can someone help me I'm trying to use ComfyUI on Windows 11. My problem more specifically is with the custom_nodes in the ComfyUI-Manager. I include an image so you can see the problem. As you can see by ComfyUI-Manager it says "Try Update" I have tried clicking this and it gives me the option to restart the server which I click on but after it restarts the same problem is still there. Also "ComfyUI Impact Pack" has a conflict warning and so do other things if I scroll down. When I click on the triangle with the exclamation mark associated with "ComfyUI Impact Pack" I get the following message "【ComfyUI Impact Pack】Conflicted Nodes (1)
SAMLoader [ComfyUI-YOLO]"

[tacit wyvern]

Why does the default inpainting model recreate the wood texture well? But dreamshaper inpainting and realisticvision inpainting cant?

Same settings/seeds. Only ckpt are changed.
Is my UI bugged?

[karmic crown]

i removed venv, relaunched ./webui.sh and in a few seconds, it was running
and i was like 🤯
i tried few more times, with same result

then i checked libraries in venv and there were pip and setup tools
everyting else was installed globally

worst part is that based on the log everything works fine, and venv is created and activated 🙂

[potent oyster]

does forge have a venv folder?

[signal sigil]

Does anyone here know why SDXL inpainting models look so bad while inpainting in Gorge webui? It blurs the edges of the inpainted area.
It has been driving me crazy ever since SDXL got inpaint models. Is there a trick that i should be doing or something? xD
[8:23 PM]
(I've also tried using it with the soft inpainting enabled and disabled. Also with both high and low mask blurring too.)

[karmic crown]

if you are using "one-click" package than no

if you are cloned from git, than it is

[potent oyster]

i need to update ultralytics, is there no venv?

[karmic crown]

wait a second i need to download it fist

[potent oyster]

shit man thanks for the effort

[potent oyster]

oh ig "environment" is the one

[karmic crown]

no

wait

you have two folders

system and webui

go to system

it has 3 folders there

go to python

and do pip install or whatever

thats it

[potent oyster]

gimme a sec testing it now

[young sierra]

hello

I'm having an issue

anyone knows how to fix it? @ me if you do

[ornate elk]

Hey install python 3.10.11 64bit
Check "add python to path"
Then delete the venv folder and relaunch the webui-user.bat

[young sierra]

how do I do the ''add python to path''

I'm struggling with that right now

[potent oyster]

there is no activate script in Scripts for some reason

[ornate elk]

what python version did you installed before?

[potent oyster]

idk how else to activate the venv

[young sierra]

3.10.6 but I uninstalled it to get 3.10.11

like you said

[karmic crown]

you do not have venv, so no activate

just run pip from the folder python

[ornate elk]

then run the installer once
and then run it again, select Modify, Next, and check "add python to environment variables"

[karmic crown]

it has its own python in that folder

[young sierra]

ok

I'll do it gimme a sec

[potent oyster]

alright im trying it

[young sierra]

I just did it now what?

[ornate elk]

delete the venv folder if there is one

if not then relaunch the webui-user,bat

[young sierra]

Installing something right now I'll let you know what happens later, thanks for helping me

😃

[ornate elk]

no problem, also dont forget to add the performance launch arsg later

depending on whats your gpu

[spring pike]

having lag spikes mid loading and when its about to be do loading , there for a short time
need help

[ornate elk]

more information please, webui version, webui-uer.bat content, gpu model, txt2img settings, cmd log

[cursive acorn]

Quick easy question: I am using img2img where I am changing the character in the image to another character with a lora. I am happy with most of the output, but the eyes are usually the spot where it fails noticeably. I normally go into Photoshop and manually edit them, but is there a way to like... inpaint just the face / eyes to improve them?

[ornate elk]

yep there is an extension exactly for that called Adetailer

[cursive acorn]

Oh perfect I'll check that out, thank you!

[ornate elk]

np!

[tame yoke]

is it possible to like merge 2 character loras and make it so the lora generates images with the characters in one photo

[ornate elk]

you mean you want two different characters based on loras in one image?

thats possible with the Regional Prompter Extensions. After installing it check this out:
#🤝｜tech-support message

[tame yoke]

anything like this for comfyui? also this only really works when they dont interact, right?…

[ornate elk]

kissing and hugging is easier, the rest not doable, idk if there is a node for comfyui

[tame yoke]

how come is it easier?

[ornate elk]

easier than other poses

sry

normal standing or sitting and not interacting is the easiest

then comes kissing, hugging, which can be done too

and everything beyond that is not easy or doable

[ionic grotto]

guys any of you have this option enabled?

[main roost]

is sd3 coming to Dreams?

[onyx solar]

Wouldn't that make it Nightmares?

[main roost]

no that was Joe Bidens debate

[onyx solar]

It's not his fault he slept through it.

[main roost]

but his eyes were too wide open to be asleep

[onyx solar]

We'll see how well you debate when you're 182 years old.

[main roost]

where am i?

i pooped my britches!!

[azure sable]

Quick question for anyone who knows about lora training...I use buckets, and all of my images are above 1024x1024, but when I look in the command window of kohya ss, I can see that it's making a bucket of 960x960 for some of my images. Is that normal or will it affect my training somehow? I never noticed it before today.

[onyx solar]

Trump didn't even attend the correct debate, though. The one he was answering the questions for was the world that exists inside his head.

It's a good thing that a presidential debate isn't the actual measure of capability in the position. It's also a good thing that the position isn't nearly as powerful as a lot of people think it is.

(But this is #🤝｜tech-support, so I won't continue to post on this topic after this; I just wanted to ensure the counterpoint was made. Life is best when there's a balance.)

[main roost]

idk what your saying...i just want to know if sd3 is in Dreams now

[onyx solar]

Oh, I think you know what I'm saying.

[main roost]

no i dont...even cnn and msnbc stated Biden is lost, he will not be the dem candidate...sorry to burst your dreamz

[onyx solar]

We'll see come November. ✌️

[main roost]

ye!

[marble frigate]

yo soo the bug is still here @ornate elk 😦

its just stuck like that

its weird if i press enter on the cmd it fixes it but i dont think its suppose to work this way

[ornate elk]

you paused the cmd by clicking on it xD

pressing space unpause it again

you can see that also at this white block

[marble frigate]

lol

thanks man

but it did fix most of the other bug stuff after i did what u told me

soo thanks for that as well

[ornate elk]

perfect, no problem 🙂

[oblique vapor]

I am sorry for a really silly question, but my python is updated to 3.10, yet I am getting an error saying I am using 3.8.2... is there any way I can rectify this issue?

[ornate elk]

you can uninstall 3.8 if not needed then run the python 3.10.11 installer again and click modify, next, and check "add python to environment variables"

then delete the venv folder and relaunch the webui-user.bat

[oblique vapor]

I think that did it... lets see if this works this time

[oblique vapor]

While I am waiting on this install, why is A1111 a web UI? Why not its own GUI application? Maybe I am dumb but I think that'd be cool

Or, does running A1111 essentially mean I run it personally and I can generate as much stuff as possible and negate the need for credits?

[young sierra]

Something strange is happening

[oblique vapor]

What's that?

[young sierra]

For some reasons

one model is working but all the others aren't

they load grey images

[fervent halo]

@ornate elk im back from vacation

but I couldnt get the video upscaler to work

[blissful nymph]

loras appear in the web folder but do not have any result when I click and add them to the prompt tired inlcuding trigger words and exluding trigger words.

[drifting pendant]

Good morning from France anyone 🇫🇷 !
I have some troubles with my installation of Automatic1111, It's kinda weird, each time I load a model into my instance, the model load into my VRAM but stays un RAM too , if load 3 models, my ram is saturated and in have to kill my process et reload anything.
I'm running Automatic1111 on "Pop!_OS" with Nvidia drivers, the latest version, on a GTX 1080 Ti, the CPU / MB / RAM is old too, but good enough :p
Does anyone have any idea how to fix this huge annoyance ?

[rare wind]

tryna run sd locally and this is popping up. using amd gpu running automatic111 any idea what i should do ?

[fossil tendon]

@karmic crown yeah so after playing with A1111 yesterday, it's doing the job on my M1 Pro but for iterating quick and experimenting the workflow is a bit slow indeed haha

[karmic crown]

@fossil tendon show me your webui-user.sh

[fossil tendon]

#!/bin/bash
#########################################################
# Uncomment and change the variables below to your need:#
#########################################################

# Install directory without trailing slash
#install_dir="/home/$(whoami)"

# Name of the subdirectory
#clone_dir="stable-diffusion-webui"

# Commandline arguments for webui.py, for example: export COMMANDLINE_ARGS="--medvram --opt-split-attention"
#export COMMANDLINE_ARGS=""
export COMMANDLINE_ARGS="--skip-torch-cuda-test --opt-sub-quad-attention --upcast-sampling --no-half-vae --use-cpu interrogate"

# python3 executable
#python_cmd="python3"

# git executable
#export GIT="git"

# python3 venv without trailing slash (defaults to ${install_dir}/${clone_dir}/venv)
#venv_dir="venv"

# script to launch to start the app
#export LAUNCH_SCRIPT="launch.py"

# install command for torch
#export TORCH_COMMAND="pip install torch==1.12.1+cu113 --extra-index-url https://download.pytorch.org/whl/cu113"
export TORCH_COMMAND="pip install --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cpu"

# Requirements file to use for stable-diffusion-webui
#export REQS_FILE="requirements_versions.txt"

# Fixed git repos
#export K_DIFFUSION_PACKAGE=""
#export GFPGAN_PACKAGE=""

# Fixed git commits
#export STABLE_DIFFUSION_COMMIT_HASH=""
#export CODEFORMER_COMMIT_HASH=""
#export BLIP_COMMIT_HASH=""

# Uncomment to enable accelerated launch
#export ACCELERATE="True"

# Uncomment to disable TCMalloc
#export NO_TCMALLOC="True"

###########################################

I didn't change anything

[karmic crown]

ok

i will 🙂

one question

have you used 1.5 of sdxl model?

[fossil tendon]

i was just thinking about it while looking at the discord's tab seeing the "sd3" and "sdxl" chans and remembered you said something about sdxl being slower than 1.5

[karmic crown]

when you install a1111 it automatically download and install default 1.5 model

[fossil tendon]

tbh i totally forgot about that detail, was just downloading models from civitai and didn't really pay attention if it was xl or 1.5 😅

so that may have caused slowness too

[karmic crown]

show me the model name

or link

even better

you can try this:

export COMMANDLINE_ARGS="--skip-torch-cuda-test --opt-sub-quad-attention --upcast-sampling --no-half-vae --medvram-sdxl --use-cpu interrogate"

this will use medvram only for sdxl models

[fossil tendon]

Oh yeah I see, it's written in the model details

[karmic crown]

yes

do not use 40 steps, use 20

it is enough

disable upscaler, upscale only images you are satisfied with

[fossil tendon]

yep this i figured it out haha

[karmic crown]

try LCM, turbo and lightning models, they work "faster", since they require less steps

[fossil tendon]

so i should focus on checkpoints/loras with base model being SD 1.5 and not XL ?

[karmic crown]

you can get great images in just 8 steps

[fossil tendon]

I generally read the recommendations on the model's page

[karmic crown]

you can use xl, but be sure to add that parameter i send

[fossil tendon]

for Realistic Vision for example they say 4-6 steps

[karmic crown]

but do not try with sde++ and 40 steps + upscaler

[fossil tendon]

alright, i just add the parameter and relaunched the webui

when applying changes to reload fully i just relaunch webui.sh right ?

[karmic crown]

close the terminal (or use CTRL+C), and run it again

it is CTRL, not COMMAND

try this model

https://civitai.com/models/200061/hephaistosnextgenxl-lcmbf16hybrid

that image is created with 9 steps as you can see

and it looks amazing

[fossil tendon]

yes i try to be mindful of the steps, it seems like some models are optimised for lower steps count

[karmic crown]

i need to go now

[fossil tendon]

thanks for your help! much appreciated

[grand whale]

#📝｜prompting-help

[ornate elk]

Exactly that

[ornate elk]

Hey, wb, what video upscaler?

[ornate elk]

Hey you need to edit the webui-user.sh
At the line commandline_args=
You need to add: --xformers --medvram-sdxl --no-half-vae
Then save and relaunch

[ornate elk]

What's your GPU?

[drifting pendant]

I'm on Linux, I have edited the webui-user.sh, but doesn't work

When I launch le instance, I see the command line args loading 😉

[ornate elk]

Then they should work

For SD on Linux you also need a large swapfile.
Especially when using sdxl models

Also 32gb ram min or an equivalent swapfile

[drifting pendant]

This is my problem

I have 16Gb of RAM / 20Gb of SWAP, and 11 Gb of VRAM

[ornate elk]

Ah okay

But 20gb swap seems okay.
What you can check too is a setting in the webui where it states how many models should stay in RAM

[fervent halo]

Oh sorry

Picture

I’m just tearing apart videos frame by frame and upscaling it image by image

[ornate elk]

Ah okay, and what upscaler doesn't work?

[drifting pendant]

Can I add --lowram ?

[ornate elk]

Yes you can try that

[fervent halo]

Ngl Idr atp

I haven’t tested it since I got back

And it’s 2am rn so I don’t wanna get back on

I’ll try again tomorrow lol

[ornate elk]

Alright okay ^^

[drifting pendant]

Sorry but I cannot find it :/

Well with the args, I cannot generate more than 2 imgs, before max out RAM and SWAP and the process crash

Without any args, I only have the problem when loading the models

I've tried "--disable-model-loading-ram-optimization" too

With args when I start the webui.sh, I'm already at 10Gb+ RAM used and 3Gb+ SWAP
Without 3.18Gb RAM / 1.56Gb SWAP

[ornate elk]

Thats normal when using an sdxl or pony model

[drifting pendant]

So I can't run args ?

[ornate elk]

It should run them

Do you have commandline_args out commented?

Like removed the #

[drifting pendant]

Yeah

For the test without the args I've commented the line to "remove" them, and I remove the "#" for enabling the line (save and relaunch too)

Maybe I'ts just my hardware, too old to run this type of workload, dunno

[ornate elk]

I have a 1080 and it worked okay with sdxl

But I was on windows with 32gb RAM

[fierce hollow]

Has the automatic111 update for SD been released?

[ornate elk]

Nope

Not yet stable

[ornate elk]

Check if the webui lists xformers at the bottom

[fierce hollow]

Please help me. I have a photo from the Internet in poor quality, is it possible to increase the image resolution using SD?

[ornate elk]

Yes thats possible

With extras tab or img2img SD upscale script

[fierce hollow]

Can I have more details, please?

[drifting pendant]

Sorry but sure to understand :/
If you are talking about what said the terminal at launch, I have a line saying : "No module 'xformers' . Proceeding without it."

[solemn moon]

hmu if any1 comes accross an xl brightness control

[ornate elk]

Then its not working

Can you screenshot the webui-user.sh commandline_args?

[ornate elk]

You have uncommented the example too?

Better remove it

Thats why it won't work

[drifting pendant]

It's like that by default, but ok

[ornate elk]

Oh okay, maybe my view on mobile screen crops it in a weird way

[drifting pendant]

Ah yeah, maybe ^^

[ornate elk]

Try to install xformers manually into the venv

go to the webui directory
source ./venv/bin/activate
cd repositories
git clone https://github.com/facebookresearch/xformers.git
cd xformers
git submodule update --init --recursive
pip install -r requirements.txt
pip install -e .

Also is your webui updated?
What version are you on?

[karmic crown]

https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/16116

FIXED 🥳

[drifting pendant]

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.
xformers 0.0.23.post1 requires torch==2.1.2, but you have torch 2.3.1 which is incompatible.
torchvision 0.16.2+cu121 requires torch==2.1.2, but you have torch 2.3.1 which is incompatible.
Successfully installed nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-nccl-cu12-2.20.5 nvidia-nvjitlink-cu12-12.5.40 nvidia-nvtx-cu12-12.1.105 torch-2.3.1 triton-2.3.1

[notice] A new release of pip is available: 23.0.1 -> 24.1.1
[notice] To update, run: pip install --upgrade pip
(venv) ai@pop-os:~/SD/stable-diffusion-webui/repositories/xformers$

Can someone kill me ? please ?

[karmic crown]

actually wait

[drifting pendant]

Too late x)

[karmic crown]

its fine

you have nvidia

on linux

[drifting pendant]

Yep

[karmic crown]

did you set torch command in webui-user.sh?

sorry i was busy with something, didnt read all of your messages

[drifting pendant]

What's that ? x)

[drifting pendant]

No worries

[karmic crown]

in your stable-diffusion-webui folder you have a file webui-user.sh

[drifting pendant]

Yes, I'm on it

[karmic crown]

lines 27 and 28 looks like this

# install command for torch
#export TORCH_COMMAND="pip install torch==1.12.1+cu113 --extra-index-url https://download.pytorch.org/whl/cu113"`

or you have something else there?

[drifting pendant]

Exactly the same

[karmic crown]

give me a second to see what it actually use for your system

[silk mantle]

Anyone know why my ponyrealism output is so bad?

This is the output I'm getting

File "C:\Users\User\Documents\Stable Diffusion\webui\venv\lib\site-packages\gradio\routes.py", line 488, in run_predict
output = await app.get_blocks().process_api(
File "C:\Users\User\Documents\Stable Diffusion\webui\venv\lib\site-packages\gradio\blocks.py", line 1434, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "C:\Users\User\Documents\Stable Diffusion\webui\venv\lib\site-packages\gradio\blocks.py", line 1335, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "C:\Users\User\Documents\Stable Diffusion\webui\venv\lib\site-packages\gradio\components\gallery.py", line 197, in postprocess
file_path = str(utils.abspath(file))
File "C:\Users\User\Documents\Stable Diffusion\webui\venv\lib\site-packages\gradio\utils.py", line 938, in abspath
if is_symlink or path == path.resolve(): # in case path couldn't be resolved
File "C:\Users\User\AppData\Local\Programs\Python\Python39\lib\pathlib.py", line 1215, in resolve
s = self._flavour.resolve(self, strict=strict)
File "C:\Users\User\AppData\Local\Programs\Python\Python39\lib\pathlib.py", line 215, in resolve
s = self._ext_to_normal(_getfinalpathname(s))
OSError: [WinError 123] The filename, directory name, or volume label syntax is incorrect: 'outputs\txt2img-images\2024-06-30\00001-475401799.png?1719742543.8466835'

[karmic crown]

ok, change that line to be like this

export TORCH_COMMAND="pip install torch torchvision --index-url https://download.pytorch.org/whl/cu118"

deactivate venv and then delete venv folder

run ./webui.sh
it will create a new venv for you

this should isntall torch 2.3.1 with cuda 118

if that does not work, we will try some older version

[karmic crown]

"?" and some other characters cannot be used in a filename

'00001-475401799.png?1719742543.8466835'

i am 99% you didnt set that yourself

[silk mantle]

Correct

The file saves regularly in the folder though, and the image quality is much lower than expected, as seen above

[karmic crown]

that is the reason for an error, but what is the cause... i really have no idea

[silk mantle]

It's def the reason for the filename error output, though would it be the reason for the image being so blotchy?

[karmic crown]

it should not be realted at all, but there might be something that is not working correctly and cause both problems

[drifting pendant]

ERROR: Cannot activate python venv, aborting...

At the execution of the webui.sh :/

[karmic crown]

give me a second

show me the result of this

echo -e "\nInstalled python versions:";for v in {10..12}; do if command -v python3."${v}" &> /dev/null;then python_version="$(python3."${v}" --version)";python_path="$(which python3."${v}")";echo -e "  \033[34m${python_version#*[[:space:]]}\033[0m: ${python_path}";fi;done;if command -v python3 &> /dev/null;then p3="$(python3 --version)";echo -e "\nDefault Python version: \033[34m${p3#*[[:space:]]}\033[0m";fi

it should be something like this

Installed python versions:
  3.10.14: /opt/homebrew/bin/python3.10
  3.12.4: /opt/homebrew/bin/python3.12

Default Python version: 3.12.4

[drifting pendant]

Installed python versions:
3.10.12: /usr/bin/python3.10

Default Python version: 3.10.12

[karmic crown]

that looks fine

[glossy wave]

where can I find the best stable diffusion? is stable diffusion 3 free?

[karmic crown]

run those one by one

which python
which python3
which pip
which pip3

and show me what you have

[silk mantle]

Anyone have any idea what might be adding the ? and numbers after the .png? I've checked the output formatting and it's regular

[karmic crown]

try to disable extensions and to run generation with just a model and all extensions disabled

if that works, than some of the plugins make a problem

[drifting pendant]

which python : Nothing
which python3 : /usr/bin/python3
which pip : Nothing
which pip3 : Nothing

[karmic crown]

also

which python3.10
which pip3.10

sorry

[drifting pendant]

which python3.10 : /usr/bin/python3.10
which pip3.10 : Nothing

[drifting pendant]

No worries ^^

[karmic crown]

ok, do this
python3 -m ensurepip --upgrade

and try
which pip3
again after that

if that works
which pip3.10
should work too

[drifting pendant]

I got "/usr/bin/python3: No module named ensurepip"

[karmic crown]

can you please reinstall python?
pip does not work for you

[drifting pendant]

Okay

[karmic crown]

https://tecadmin.net/how-to-install-python3-on-popos/

you can use ppa:deadsnakes/ppa

i use that on ubuntu

[silk mantle]

Hmm.. Getting the same issue w all extensions disabled

[karmic crown]

try another model

[silk mantle]

I did, tried a few and all gave the same issue

[karmic crown]

your webui might be corruptet, try to update

[silk mantle]

Do you think I need to try a full 1111 reinstall? Would suck

[karmic crown]

just do git pull

[silk mantle]

Funny bc I just did a git pull right before this session

So maybe the update itself is the corruption?

[karmic crown]

ok show me git status

[silk mantle]

Did a git pull via the user ui bat

[silk mantle]

Sorry for the noob level knowledge here, where do I find that?

[karmic crown]

you need to open CMD, go to your stablediffusion folder

and type git status there

[drifting pendant]

I cannot remove the current Python

[karmic crown]

blah...

[drifting pendant]

Well I want to trash all this and restart form scratch

[karmic crown]

how about
sudo apt install --reinstall python3

[silk mantle]

Hmm... I may have messed something up...

PS C:\Users\User\Documents\Stable Diffusion\webui> git status
On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git restore <file>..." to discard changes in working directory)
modified: webui-user.bat

no changes added to commit (use "git add" and/or "git commit -a")
PS C:\Users\User\Documents\Stable Diffusion\webui>

[karmic crown]

looks good

only webui-user.bat is changed

everything else is the latest version

[silk mantle]

Oh ahaha maybe I didn't mess it up then

[karmic crown]

that is really strange

[silk mantle]

I'll show you what the inside of my webui-user.bat looks like

[silk mantle]

@echo off

set PYTHON=
set GIT=
set VENV_DIR=
set COMMANDLINE_ARGS=--autolaunch --medvram --xformers --no-half --disable-nan-check

call git pull
call webui.bat

[drifting pendant]

Reinstalled but nothing when I type : which pip3.10 and which pip3

[karmic crown]

add --disable-all-extensions (you will remove it later)

and run again

[karmic crown]

python3 -m pip --version

is this works?

[drifting pendant]

/usr/bin/python3: No module named pip

[karmic crown]

python3 -m pip install --upgrade pip

[drifting pendant]

It's the same : /usr/bin/python3: No module named pip
IDK Why

[karmic crown]

i see now, that was stupid sugestion

python3 -m ensurepip --version

is this working

[drifting pendant]

Sadly not, same as last time : /usr/bin/python3: No module named ensurepip

[karmic crown]

ok

forget python3.10

[karmic crown]

install 3.11 from deadsnakes/ppa

that will have all the things you need

[silk mantle]

Same issue

[karmic crown]

when you install that and confirm that both
python3.11 and pip3.11 works
you will need to add this to your webui-user.sh
python_cmd="python3.11"

its in line 16

[karmic crown]

delete venv and rerun webui-user.bat
it will create new one
any change?

[drifting pendant]

I've installed the version 3.11 and when I "python3 --version" : "Python 3.10.12"

[karmic crown]

its ok

[silk mantle]

ERROR: Could not find a version that satisfies the requirement torch==2.1.2 (from versions: 2.2.0, 2.2.0+cu121, 2.2.1, 2.2.1+cu121, 2.2.2, 2.2.2+cu121, 2.3.0, 2.3.0+cu121, 2.3.1, 2.3.1+cu121)
ERROR: No matching distribution found for torch==2.1.2

[karmic crown]

python3.11 --version

[drifting pendant]

Python 3.11.9 🙂

[karmic crown]

you are using pyton3.12

you need python 3.10

[silk mantle]

How do I downgrade that?

[karmic crown]

https://github.com/CS1o/Stable-Diffusion-Info/wiki/Installation-Guides

just download the right version

[silk mantle]

So I don't have to deacticate 3.12?

[karmic crown]

and pip3.11 --version (please tell me that it works)

[karmic crown]

i have all versions from 3.8 -3.12 🙂

i need them for diffrent tests

[drifting pendant]

ai@pop-os:~/SD/stable-diffusion-webui$ pip3.11 --version
Command 'pip3.11' not found, did you mean:
command 'pip3.10' from deb python3-pip (22.0.2+dfsg-1ubuntu0.3)
Try: sudo apt install <deb name>
ai@pop-os:~/SD/stable-diffusion-webui$ pip3.10 --version
Command 'pip3.10' not found, but can be installed with:
sudo apt install python3-pip

Well... If that's not "F" up

[karmic crown]

what?!

run this

sudo apt install python3-pip

[ornate elk]

whats your gpu?

never use --disable-nan-check

[drifting pendant]

After the install :
ai@pop-os:~/SD/stable-diffusion-webui$ pip3.11 --version
Command 'pip3.11' not found, did you mean:
command 'pip3.10' from deb python3-pip (22.0.2+dfsg-1ubuntu0.3)
Try: sudo apt install <deb name>
ai@pop-os:~/SD/stable-diffusion-webui$ pip3.10 --version
pip 22.0.2 from /usr/lib/python3/dist-packages/pip (python 3.10)
ai@pop-os:~/SD/stable-diffusion-webui$

[karmic crown]

note to self: "some c.... distributions have a f.... python3-pip package" 😱

[silk mantle]

3080 12gb

[silk mantle]

Getting the same error after installing 3.10

[ornate elk]

then remove --no-half --disable-nan-check and change --medvram to --medvram-sdxl

[karmic crown]

i told him to remove venv and now cannot install torch 2.1.2

so it he is obviously on 3.12

oh wait

sorry

[karmic crown]

didnt notice this

[karmic crown]

@ornate elk remember this, when someone else start pulling hair off becuse he cannot install pip

@drifting pendant try again now

but without 3.11 in webui-user.sh

just commend that line (add # at the begining)

we finally fixed your 3.10 installation

[silk mantle]

Output I'm getting:

The Python error, then:

[notice] A new release of pip is available: 24.0 -> 24.1.1
[notice] To update, run: C:\Users\User\Documents\Stable Diffusion\webui\venv\Scripts\python.exe -m pip install --upgrade pip
Traceback (most recent call last):
File "C:\Users\User\Documents\Stable Diffusion\webui\launch.py", line 48, in <module>
main()
File "C:\Users\User\Documents\Stable Diffusion\webui\launch.py", line 39, in main
prepare_environment()
File "C:\Users\User\Documents\Stable Diffusion\webui\modules\launch_utils.py", line 380, in prepare_environment
run(f'"{python}" -m {torch_command}', "Installing torch and torchvision", "Couldn't install torch", live=True)
File "C:\Users\User\Documents\Stable Diffusion\webui\modules\launch_utils.py", line 115, in run
raise RuntimeError("\n".join(error_bits))
RuntimeError: Couldn't install torch.
Command: "C:\Users\User\Documents\Stable Diffusion\webui\venv\Scripts\python.exe" -m pip install torch==2.1.2 torchvision==0.16.2 --extra-index-url https://download.pytorch.org/whl/cu121
Error code: 1
Press any key to continue . . .

[drifting pendant]

Maybe it was when I try this, everything was broken

[karmic crown]

sudo apt install python3.10 python3.10-venv python3.10-dev
do this first

[ornate elk]

Please share the full cmd log

[drifting pendant]

ERROR: Cannot activate python venv, aborting...

[karmic crown]

ok

this is for ubuntu

obviously packages are not the same

rerun ./webui.sh

[silk mantle]

Already up to date.
venv "C:\Users\User\Documents\Stable Diffusion\webui\venv\Scripts\Python.exe"

INCOMPATIBLE PYTHON VERSION

This program is tested with 3.10.6 Python, but you have 3.12.3.
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-3106/

Alternatively, use a binary release of WebUI: https://github.com/AUTOMATIC1111/stable-diffusion-webui/releases/tag/v1.0.0-pre

Use --skip-python-version-check to suppress this warning.

1/2

[silk mantle]

Python 3.12.3 (tags/v3.12.3:f6650f9, Apr 9 2024, 14:05:25) [MSC v.1938 64 bit (AMD64)]
Version: v1.9.4
Commit hash: feee37d75f1b168768014e4634dcb156ee649c05
Installing torch and torchvision
Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/cu121
ERROR: Could not find a version that satisfies the requirement torch==2.1.2 (from versions: 2.2.0, 2.2.0+cu121, 2.2.1, 2.2.1+cu121, 2.2.2, 2.2.2+cu121, 2.3.0, 2.3.0+cu121, 2.3.1, 2.3.1+cu121)
ERROR: No matching distribution found for torch==2.1.2

[notice] A new release of pip is available: 24.0 -> 24.1.1
[notice] To update, run: C:\Users\User\Documents\Stable Diffusion\webui\venv\Scripts\python.exe -m pip install --upgrade pip
Traceback (most recent call last):
File "C:\Users\User\Documents\Stable Diffusion\webui\launch.py", line 48, in <module>
main()
File "C:\Users\User\Documents\Stable Diffusion\webui\launch.py", line 39, in main
prepare_environment()
File "C:\Users\User\Documents\Stable Diffusion\webui\modules\launch_utils.py", line 380, in prepare_environment
run(f'"{python}" -m {torch_command}', "Installing torch and torchvision", "Couldn't install torch", live=True)
File "C:\Users\User\Documents\Stable Diffusion\webui\modules\launch_utils.py", line 115, in run
raise RuntimeError("\n".join(error_bits))
RuntimeError: Couldn't install torch.
Command: "C:\Users\User\Documents\Stable Diffusion\webui\venv\Scripts\python.exe" -m pip install torch==2.1.2 torchvision==0.16.2 --extra-index-url https://download.pytorch.org/whl/cu121
Error code: 1
Press any key to continue . . .

2/2

[karmic crown]

i misread your message, sorry, i tough it is about the package

[ornate elk]

Uninstall Python 3.12, its not compatible.
Then install python 3.10.11 64bit
Then delete the venv folder
Then relaunch the webui-user.bat

[karmic crown]

@drifting pendant try
sudo apt install virtualenv

sudo apt install python3-virtualenv

one of those should work

[drifting pendant]

Installed this one : Same

[drifting pendant]

Alerady installed, Same, I try to reinstall it RN

Same

[karmic crown]

can you please rename folder with stablediffusion to anything else like sd-bkp

and do this

wget -q https://raw.githubusercontent.com/AUTOMATIC1111/stable-diffusion-webui/master/webui.sh

./webui.sh

now you have all python components including venv and pip

so try again

[silk mantle]

Getting this now: Already up to date.
Couldn't launch python

exit code: 9009

stderr:
'python' is not recognized as an internal or external command,
operable program or batch file.

Launch unsuccessful. Exiting.
Press any key to continue . . .

[karmic crown]

i ❤️ those spammers

[drifting pendant]

Repport ?

[drifting pendant]

I'm making a backup of my models and I do that ASAP thx

[ornate elk]

Run the Python 3.10.11 installer again.
Select Modifiy
Click next,
check "add python to environment variable's"
Then relaunch the webui-user.bat

[drifting pendant]

I'm runing it I have "No module 'xformers'. Proceeding without it."

I can run SD again

[karmic crown]

ok

so now you need xformers

but..

wait now

ls -l venv/lib/python3.10/site-packages

show me what you have there

[karmic crown]

transformers-4.30.2 is there

do not install anything else

you just need different paramers in webui-user.sh

line 13
#export COMMANDLINE_ARGS=""

it needs to be
export COMMANDLINE_ARGS="--xformers"

start with that

what is your GPU?

[potent oyster]

I have a bunch of images i generated that i want to use for training, is there any way of extracting the prompt from the meta data and putting it in a txt?

[drifting pendant]

A humble 1080Ti

[karmic crown]

@ornate elk what params should be used for 1080Ti? --xformers --medvram? or what

i only know parameters for macs by heart 🙂

sorry

[potent oyster]

had a 1070 a while ago, those args worked for me

id definitely use them

[karmic crown]

@drifting pendant then try those two

should get rid of that warning

[drifting pendant]

--xformers --no-half-vae --medvram-sdxl

I think

[karmic crown]

you can try that

10xx does not need --no-half as 16xx

so that make sense

[drifting pendant]

I've just RTFM of @ornate elk

[drifting pendant]

And yes

UMMAGAD : Applying attention optimization: xformers... done.

[karmic crown]

at least i learned something - python packages for popos are c... 😂

[crude trellis]

following zluda instalation proccess on w11

just wanted to confirm if web-user.bat commands are as planned

is everything ok?

[karmic crown]

looks fine, but you can remove space between = and --

[crude trellis]

ok

[drifting pendant]

me too, thanks for all you have done !

[karmic crown]

looks fine

[drifting pendant]

@karmic crown @ornate elk If I can help for anything, tell me guys.

[crude trellis]

btw, I've tried following this tutorial before some months ago, but ended up giving up halfway

[silk mantle]

60% of my problem has been solved, thank you! The filenaming issue is gone. However, I'm still getting the issue of images looking wrong. I've attached an example image from civitai, and my recreation using png info, which is coming out just looking underdeveloped?

[crude trellis]

is there any way I can confirm that AMD HIP SDK 5.7 is correctly installed?

or would you guys just recommend to delete it and do a fresh install or smt

[crude trellis]

cute

[karmic crown]

better go to #📝｜prompting-help for further help

[drifting pendant]

One last question, what model can I run on this GPU W/O corrupting anything ? Like SDXL / SD1.5 / any Pony ?
Does the seize of the model in GB matter ?

[ornate elk]

If you installed it, and added it to path it should work.
Also for your card you need the custom rocm library files
And a PC restart after that

[karmic crown]

always use the smallest model first

usually it is marked as pruned fp16

if that works great

smaller filesizes, less memory usage...

pruned fp16 1.5 models are around 2GB, and sdxl around 6.5GB

[drifting pendant]

Seems logic, and this what I've observed, I don't want to make any mistakes anymore, thanks

[karmic crown]

you can try using lightning and turbo based models to get images with less steps, which be "faster" since it will take less time

[drifting pendant]

One other thing, If I play with SDXL I have to lower the render resolution, for VRAM sake

[ornate elk]

To get the same image like on civitai is not alwaye possible but make sure you do it like this:
On the image in civitai click on the blue copy prompt Settings button.
Then paste that into the positive prompt.
After that Click on the blue/white arrow at the right.
It should now apply everything automaticly.
Then generate.

[silk mantle]

Thank you!

[karmic crown]

he was using the different version of a model, we figured that out in #📝｜prompting-help

[drifting pendant]

Well, everything is working like a charm, thanks again !

[karmic crown]

it could be faster it you os was normal 😂

[crude trellis]

the problem is I that I installed it months ago

just wanted to check out if everything is fine

I don't even remeber if I uninstalled it or not

[drifting pendant]

All "because" of this video : https://www.youtube.com/watch?v=Wjrdr0NU4Sk
But thanks to Him I was able to start something

[karmic crown]

i would ALWAYS choose ubuntu for server

[ornate elk]

then check the system control pannel

if you have this

[crude trellis]

ok

[drifting pendant]

I'm more a debian guy, but I understand in this case

BTW I have some "CPU leak" (wierd RAM usage / overload) when I load models, but this is wayyyyy better than before, still strange

[karmic crown]

try --disable-model-loading-ram-optimization

[drifting pendant]

I've mistaken, the load seems stable, "high", but stable, last time I've tried this it was on a server with 256Go of RAM, so... 10Gigs of load was kinda nothing

[ornate elk]

@crude trellis let me know hows it going

[crude trellis]

sure

im at control pannel rn

where should I go

[ornate elk]

click on "uninstall a programm"

then you have the full programm list

[crude trellis]

yup

they're here

now I should check if zluda is in PATH right?

how do I confirm that?

[ornate elk]

by looking into the path environment settings. the step is listed in the guide

@crude trellis

[crude trellis]

@ornate elk

like this?

[ornate elk]

no

click on path

and then on edit

then on new

[crude trellis]

here?

[ornate elk]

yes

[crude trellis]

like this then?

[ornate elk]

yes

then also add %HIP_PATH%bin

[crude trellis]

my gpu is a rx6750xt

[ornate elk]

no wait

[crude trellis]

so I should follow step 4

right?

[ornate elk]

you have clicked on the PATH at the user variables

you need to add them to the system variables path

[crude trellis]

here?

[ornate elk]

no

just on Path

[crude trellis]

strange

[ornate elk]

scroll down a bit

[crude trellis]

there is no Path on system variables

[ornate elk]

oh really?

[crude trellis]

yep

idk if I deleted that by mistake or smt

[ornate elk]

hmm strange

okay than let them in the User environments

it should work

[crude trellis]

ok

going for step 4 now

done

gonna restart my PC now and see if it worked

(lets hope it does)

done

@ornate elk

[ornate elk]

wrong python version

uninstall python 3.12
install python 3.10.11 64bit
check "add python to path"
delete the venv folder
relaunch tghe webui-user.bat

[crude trellis]

ok

[crude trellis]

where can I check "add python to path"?

[ornate elk]

if its not there, relaunch the python installer, click modify, next, check "add python to system veriables"

[shrewd rune]

Is there a mobile site where you can generate images.

[ornate elk]

next

[crude trellis]

oh

there it is

deleted venv

lets see how it goes

does this part usually take this long?

-_-

[timber holly]

does anyone know how SD interprets it if I mistyped a lora prompt with a comma? i.e. <lora:model:0,6>

[karmic crown]

do you have COMMANDLINE_ARGS= correctly set?

[karmic crown]

did you run webui-user.bat or webui.bat?

[crude trellis]

webui-user.bat

[karmic crown]

you need to run webui-user.bat

hmm

strange

we need that switch for macs, but for zluda it should not be there

[crude trellis]

@ornate elk do you think this could be related to not having Path in System variables?

[ornate elk]

did you git cloned the webui today?

because yours is still called directml at the end

[crude trellis]

nope

its been there for some months

[ornate elk]

then please do a fresh install

[crude trellis]

ok

[craggy oxide]

i know its been a long time but this helped me so much, i've got my hair falling off! thanks!

[tacit wyvern]

Anyone knows how to fix the halo/blur edges effect for inpainting with Automatic1111?

[ruby yoke]

what are your settings

[crude trellis]

@ornate elk it opened!!

so does this means its working 100% or do I need to do some more stuff

[ornate elk]

nice! now you only need to gen the first image
it will take 15-40 minutes
check the cmd for any error

[ornate elk]

ohhh

download this model:
https://civitai.com/models/4384/dreamshaper
put it into the models/stable-diffusion folder

then delete the 1.5 ema pruned model
then relaunch

[crude trellis]

ok

[crude trellis]

after generating the first image, how long does it usually take to gen?

[ornate elk]

seconds

maybe 1 minute with sdxl but seconds with 1.5 models

then it depends on the extensions used and resolution too

[crude trellis]

btw would 500x700 be considered a large resolution?

that would take longer

[ornate elk]

no thats a normal resolution for 1.5 models (512x768) and would take seconds

[crude trellis]

great

trying to generate something now using dreamshaper

last cmd message was this

does this means its on the generation proccess?

[ornate elk]

yes, it wont show anything until the zluda load is done

the generate button should be grayed out

then its working

just let it load until you get an image 🙂

[crude trellis]

its working then

[ornate elk]

can you screenshot the whole cmd too?

just to be sure there is no error before dogetexx

[ornate elk]

perfect

[crude trellis]

thats really great to hear, directml used to take minutes and I always had some vram problems

like messages saying that I dont got enough vram

but i have 12gb

[ornate elk]

yes that wont happen with zluda that fast

less vram usage and faster generation speed

and if you install the tiled vae extension later you can also upscale without problems

[fossil tendon]

guys, out of curiosity, what's the average budget for a good pc to do only SD? I am a Apple guy so I have absolutely no idea about pc hardware and how much it costs

[fading heron]

I followed the directions on Github to install Stable Diffusion and everything appears to have installed properly, except the command prompt dissappears after running the webui-user.bat file. I see it perform nummerous functions on the command line. then closes. I'm trying to run on Window 11 with an AMD RX 7900 XT. I would really appreaciate any help on opening and running Stable Diffusion.

[ornate elk]

Hey I have the same card, did you followed my guide?

[fading heron]

I'm not sure, I would appreciate a link. I used the info on Github.

Looking at the terminal, it may be because I am using Python 3.12.4

Do I need to downgrade python, or will it run on the current version?

[ornate elk]

You need to downgrade

3.10.11 64bit is good

Here is the guide:
https://github.com/CS1o/Stable-Diffusion-Info/wiki/Installation-Guides#amd-automatic1111-with-zluda

[fading heron]

Thank you so much. I will downgrade my python then try again.

[crude trellis]

@ornate elk it worked!!!

thank you so much

seriously

you're amazing

if I install models and loras

I'll have to restart SD and wait 40 mins again so i can use them?

[tacit wyvern]

I fixed it, by using a normal base model + controlnet inpainting.
Was using an inpainting model, and it keeps creating halo/blurs around the inpainted mask.

Only sd 1.5 inpainting worked. But controlnet fixed that now.

[crude trellis]

or will they be available instantly?

[ornate elk]

No they will be available instantly

No problem, have fun 🙂

[crude trellis]

greaaat

ty

[ornate elk]

@fading heron let me know if it worked or if you need help

[blissful nymph]

people talk about sd 1.5 and 1.0 can someone help me understand what thats all about? I just downloaded the latest web ui version. Should I be doing something else for better results ? and if so how

[ornate elk]

Hey, you need to download community models from Civitai.com
These give much better results than the default 1.5 model

For example try Dreamshaper v8

[blissful nymph]

ty for the info. One other question @ornate elk I downloaded a lora but when i inset it <> into the prompt and or use the trigger words nothing seems to happen. What might I be doing wrong?

-- for further context I installed the loras in stable-diffusion-webui -> models / Lora

[ornate elk]

the lora needs to be compatible with the model

on civitai you always see the "based on" in the info box, there you see if its for 1.5 or sdxl or pony

[blissful nymph]

ty @ornate elk

[fossil tendon]

@karmic crown in case you also know about pc hardware and not only macs 🙂

[karmic crown]

at least nvidia 3070, ideally 4070 or better

[karmic crown]

i used to work in pc shop 20 years ago, but i switched to web and software development 🙂

[karmic crown]

maybe it is better ask in #💬｜general-chat since some people who are there do not come here 🙂

[fossil tendon]

sure, will do! I thought it was more of a tech related topic that's why i posted here

[ornate elk]

only for SD can mean anything. you can use SD with a 4gb gpu but its not great

best is to go for a gpu with 12gb vram or more from nvidia

[karmic crown]

he has working m1, so nothing with less with 12gb or he will waste money

[fossil tendon]

yes, need something faster than my m1

[karmic crown]

i guess 3070 with 12GB would be enough to feel a real difference

[ornate elk]

3070 has 8gb

[karmic crown]

few times faster

which one is with 12GB? 4070?

but even 3070 with 8GB is much faster than that m1

https://vladmandic.github.io/sd-extension-system-info/pages/benchmark.html

find what you can tolerate

[fossil tendon]

basically i don't want to wait minutes when generating regular images (by regular i mean without upscaling)

to have a fluid worklow when working on images

[karmic crown]

this is fine

its few times faster than my m3 pro

3070 Ti

this means about 4 seconds per 40 steps 🙂

[fossil tendon]

that sounds about right

and regarding the rest of the pc? cpu/ram? any recommended brands?

[karmic crown]

i guess 16GB is reasonable

if you do not plan to use that for many other things

cpu, i guess some ryzen

i would not waste money on intel

[ornate elk]

definitly amd ryzen

[karmic crown]

i guess ryzen 5 would be enough, right?

[fossil tendon]

nope, the m1 pro is for software development (i do iOS dev) and I have my synology server, I do not need another machine at all, only for SD

cool, so amd ryzen, 16gb ram, 370ti

[karmic crown]

then 16GB it is

[fossil tendon]

i will search for that and see what i find, maybe will post on general later to have other people opinion

[karmic crown]

@fossil tendon pleas dm me, i have to ask you something aboug ios emulator in xcode

[light prism]

Hey i am wondering if there are any ideas to how to modify images with the prompts by describing the operation rather than the output. For example using a image of a robot and a prompt "add an arm" rather then the detailed original prompt extended with the total number of arms "high tech robot, red, photorealistic ..., three arms" and so on. Are there models that are capable or this? In specific cases i got the desired output, but most often the outputs seamed random, especially when asking for "remove XY" it adds the XY and so on

[crude trellis]

only 3 of my loras are showing up fsr

I have 20+

usually they all are visible

[ornate elk]

Switch to a compatible model

Then reload the loras

[low basin]

Got a noob Python question- For those of you running Windows and who have multiple installs of Python - How are you handling the system paths for each instance/version? Due to application requirements I've now amassed 4 separate versions and it has become a proper PITA to have to dig through every program and manually set a "python=" variable due to most applications and scripts just assuming "python=python" will pull the correct version.

Is there a way in System Variables to set paths to a specific python path/install that only is traced if executed in a certain directory? So depending on what folder/directory you attempt to execute python from it will grab a certain path/version?

[ruby yoke]

if it's me, i only have the path to the version of python i'm using configured. and if i have to use a different version, i'm going to change the pathstatement to that version.

[low basin]

Are you saying you change the path statement in the system/user variables or in the script's variables? If the former, that's basically what I'm doing now but when bouncing back and fourth between scripts it gets cumbersome to keep swapping it

[near stratus]

hello, for some reason when I try to create a new hypernetwork, and train it I receive this error:

Applying attention optimization: xformers... done.
*** Error completing request
*** Arguments: ('task(o25gm3oq98g3eiv)', 'LTAI', '0.0001', 1, 1, 'C:\\Users\\anriq\\Downloads\\AI_test\\LT\\treinotxt', 'textual_inversion', 512, 512, False, 2500, 'disabled', '0.1', True, 0, 'once', False, 250, 500, 'hypernetwork.txt', False, '', '', 20, 'DPM++ 2M', 7, -1, 512, 512) {}
    Traceback (most recent call last):
      File "C:\Users\anriq\Desktop\StableDiffusion\stable-diffusion-webui\modules\call_queue.py", line 57, in f
        res = list(func(*args, **kwargs))
      File "C:\Users\anriq\Desktop\StableDiffusion\stable-diffusion-webui\modules\call_queue.py", line 36, in f
        res = func(*args, **kwargs)
      File "C:\Users\anriq\Desktop\StableDiffusion\stable-diffusion-webui\modules\hypernetworks\ui.py", line 25, in train_hypernetwork
        hypernetwork, filename = modules.hypernetworks.hypernetwork.train_hypernetwork(*args)
      File "C:\Users\anriq\Desktop\StableDiffusion\stable-diffusion-webui\modules\hypernetworks\hypernetwork.py", line 483, in train_hypernetwork
        hypernetwork.load(path)
      File "C:\Users\anriq\Desktop\StableDiffusion\stable-diffusion-webui\modules\hypernetworks\hypernetwork.py", line 250, in load
        self.layer_structure = state_dict.get('layer_structure', [1, 2, 1])
    AttributeError: 'NoneType' object has no attribute 'get'

I can only train old hypernetworks models, someone can help me ?

[ruby yoke]

i would change it in my system variables. but that's just me, personally. mostly i only use one version, so using other verions is going to be a rare thing

[azure magnet]

Hi, why my pc use the tiled version of vae, even if i have enought ram?

In automatic1111

[rare wind]

rx6600

[storm sail]

anyone here familiar with openvino?

[river bough]

Hey, I've also had the same issue as Frusto where some LoRA models were not visible, but I checked an option and now all the LoRA models are visible. Do you know which option that is to desactivate it again?

Nvm i already found the option lol

[ornate elk]

Follow my Automatic1111 with ZLUDA guide from the pinned messages of this channel

[rare wind]

works flawlessly even tried different models and no issues. thanks heaps !!!

[white thorn]

Hey, is A1111 still on Python 3.10.6 ? or is it possible to update to .. 3.11 ?

[azure magnet]

Hi, why my pc use the tiled version of vae, even if i have enough ram with automatic1111?

[light prism]

Since i didnt get any response ill repost my question - any kind of hints are appreciated: Hey i am wondering if there are any ideas to how to modify images with the prompts by describing the operation rather than the output. For example using a image of a robot and a prompt "add an arm" rather then the detailed original prompt extended with the total number of arms "high tech robot, red, photorealistic ..., three arms" and so on. Are there models that are capable or this? In specific cases i got the desired output, but most often the outputs seamed random, especially when asking for "remove XY" it adds the XY and so on

[ornate elk]

Perfect, no problem 🙂

[ornate elk]

Where do you see that it tiles the VAE?

[ornate elk]

Yes A1111 works still the best with 3.10.11 64bit

[azure magnet]

I import the image in comfyui

[ornate elk]

Hmm okay

RAM has nothing to do with Vae

A VAE uses vram

[azure magnet]

i have 12gb of vram

[ornate elk]

And do you have a VAE tiling node? Or where do you see that it gets tiled ?

[azure magnet]

When i import the image that i have generated make with automatic1111, in comfyui

Could be here the problem?

Could i fix it?

[ornate elk]

Thats not tiled vae

I think comfyui automatically tiles the VAE not matter what

[azure magnet]

Ok thanks a lot

[runic pike]

hi everyone, could anyone can tell me that how deploy the stable diffusion 3 in colab with colab GPU but the API? thank you very much.

[white thorn]

https://tenor.com/view/think-about-it-use-your-brain-use-the-brain-think-brain-gif-7914082

[craggy surge]

getting this issue on a1111

[ornate elk]

What's the full cmd log?

[craggy surge]

Well I tried running the update.bat and now its installing some requirements

when I tried windows run

again

[silk mantle]

Hey, I'm getting an issue in the DreamBooth extension (full log in link below) that basically amounts to "cannot load because keys are missing", and then at the very end, "Please make sure to pass low_cpu_mem_usage=False and device_map=None if you want to randomly initialize those weights or else make sure your checkpoint file is correct.'."

I've tried the solutions I've found online, with multiple models, but nothing has worked.

I'm currently trying with the HelloWorld and Juggernaut models.

I previously got the "diffusion_pytorch_model" missing in working folder error, and got past that by downloading the pytorch model from the huggingface repo. Not sure whether it has anything to do with this?

Tried posting in the DreamBooth extension discord but I'm not sure whether it's active.

Full log: https://justpaste.it/7pv72

[ornate elk]

is it fixed now?

[ornate elk]

dont use Dreambooth, use Kohya_ss for training models or Loras

or OneTrainer

[craggy surge]

downloading stuff

it worked

[cinder ermine]

Im trying to get onetrainer working, i seem to be getting repeated errors when trying to use a non official model

tried using a file saved locally and a huggingface hosted file

what am i doing wrong? I keep getting errors when i start training

[ruby yoke]

can you screen shot and post the errors?

[fading heron]

hi, hope your doing good.

I followed the guide you sent me and downgraded my python to 3.10.11, but now I am getting a no python found error.

I tried reinstalling python and made sure that the "add to Path" box was checked. Do you have any idea why this could be happening. Sorry to bother you.

[ornate elk]

Have you deleted the venv again too?

[fading heron]

no, I'm not sure I know what that is. Actually I'm sure I don't know what that is.

[fair oxide]

@knotty turret welcome, try here with tech questions

[knotty turret]

ok

[fading heron]

no, I'm not sure I know what that is. Actually I'm sure I don't know what that is.

[fading heron]

I don't have that directory in my c:\SD-ZLUDA directory

[fair oxide]

thought you were running auto1111, idk about zluda, might as well wait for cs1o

[fading heron]

Hey, thanks for trying.

[ornate elk]

Go into the stable-diffusion-webui-amdgpu folder and delete the venv folder

[fading heron]

I'm in that directory, but that folder does not exist.

[fading heron]

I'm in that directory, but that folder does not exist.

[elfin wyvern]

Hey! Quick question (sorry if it was answered already - it probably was): In ComfyUI, do I have to tag a lora in positive prompt if im using a lora loader node?

[ornate elk]

Then relaunch the webui-user.bat

If it wont work, relaunch the python installer, select modify, next, ans check "add python to environment variable's"

[fading heron]

I verified that the "add python to ..." was checked when I ran the installer. I ran webui with the arguments

--ckpt-dir "Path/to/models/stable-diffusion" --lora-dir "Path/to/models/lora" --controlnet-dir "..." --embeddings-dir "..."

[ruby yoke]

did you add python to your system path?

[fading heron]

result was couldn't launch python exit code 9009

[fading heron]

it should be, could you tell me how to verify?

[ruby yoke]

windows or mac?

[fading heron]

windows 11

[fading heron]

windows 11

[ruby yoke]

read through this https://www.c-sharpcorner.com/article/how-to-addedit-path-environment-variable-in-windows-11/

[fading heron]

i will, thank you

[ruby yoke]

welcome

[fading heron]

Thank you so much, I had to add the python directory to PATH. It

It's downloading and installing. Fingers crossed

[fading heron]

It was the Path variable. Sorry for the trouble. Thank you so much for your help.

[ruby yoke]

no trouble at all 🙂 happy it's working for you now.

[fading heron]

It's working!!!

[tacit lynx]

Nice to meet you,
I'm currently using stable diffusion forge, is there an extension that separates Lora's description area?
I've used forge-couple, but it doesn't work well for very complex things. If there are other extensions, I would like to know about them.

[upbeat crane]

i've been having decent luck with cutting down on some prompt bleeding and lora control by simply utilizing some of the lora tags within the adetailer prompts....lots of adetailer models to sift through on civit. just a thought. havent tried much in the way of extentions but this is worth checking out:

https://www.reddit.com/r/StableDiffusion/comments/1aqlvi0/psa_dont_ignore_the_sdwebuiloractl_extension_for/

i would love to hear some feedback on LoraCTL if you end up working with it, good luck mate

[graceful crag]

Hello

[upbeat crane]

howdy drake

[dusty vine]

Is fine-tuning not available through the API? https://platform.stability.ai/sandbox/fine-tuning

[silk mantle]

Thanks, I'll give that a go!

[fickle charm]

Hi all, new here, I am looking to purchase a gtx 1660 super 6gb GPU. Have the known issues of black/green screen output in SD been fixed or not?

[fickle charm]

I can still cancel my order, so any help/response on above will be greatly appreciated

[fiery idol]

Hi there! I'm unable to cancel my subscription to Stable Audio, I submitted a ticket almost a month ago and no answer. When I try to access my profile page or the pricing page there's a loading spinning wheel but nothing else happens.....any help???

[ornate elk]

A Gtx1660 is the worst GPU for using SD, as this GPU doesn't support half float point (fp16)
Everything will take longer and needs more vram

But the black output is no problem anymore

[quaint egret]

i am trying to install it and i cant get it to work... can someone help me please

[vocal burrow]

cf pinned guides

[chilly harbor]

Hello! Iam trying to install SD and had first to install Swarm, right? so i installed it but i cant find the folder

Oh nevermind its working now

[chilly harbor]

SD is almost working but it says "Backends are still loading on the server..." and (Git failed to load)

I did all the steps from Reddit

[ornate elk]

You don't need Swarm to use SD, but you can. Its one of many stable diffusion webuis

[chilly harbor]

It says i didnt clone from git or something like that

Do you recommend using something else?

[ornate elk]

Yea if your new start with Automatic1111.
You'll find the installation guide in the first link from the pinned messages.

[chilly harbor]

Do i have to install Python?

Because when i run it, it says Couldn't launch python

Couldn't launch python

exit code: 9009

stderr:
Python was not found; run without arguments to install from the Microsoft Store, or disable this shortcut from Settings > Manage App Execution Aliases.

Launch unsuccessful. Exiting.
Press any key to continue . . .

[ornate elk]

Yes

Python 3.10.11 64bit

And Git

[chilly harbor]

Ohh i have 3.12

[ornate elk]

Thats not compatible, you can uninstall it

[chilly harbor]

I have installed Python 3.10.11 64bit now and Git but its still not working

Python was not found; run without arguments to install from the Microsoft Store, or disable this shortcut from Settings > Manage App Execution Aliases.

What does this mean?

It does say amd, but i dont have amd

[ornate elk]

Thats normal, amd is not related to amd brand in this case

[ornate elk]

Run the installer again.
Click Modify, Click next, check "add python to environment variables"

[chilly harbor]

Thanks : ) its working now

[silk mantle]

Which folder would you put an embedding saved as a safetensors in? I have it in the regular SD folder, but the output is coming up as plain grey

[chilly harbor]

Now it says this: AttributeError: 'NoneType' object has no attribute 'lowvram'

[silk mantle]

Example of output. Here's the PNG info from it:
parameters

AIMe man taking a selfie on a sidewalk
Steps: 20, Sampler: DPM++ 2M, Schedule type: Karras, CFG scale: 2.5, Seed: 331226645, Size: 832x1216, Model hash: 5b10498158, Model: AIMe, Clip skip: 2, Downcast alphas_cumprod: True, ComfyUI Workflows: {}, Version: v1.9.4

[chilly harbor]

Its now working by removing " --no-half-vae"

[ornate elk]

Delete the 1.5 EMA pruned model from models/stable-diffusion and instead download the Dreamshaper v8 model from Civitai.com.
Put that into models/stable-diffusion folder ans then relaunch.
https://civitai.com/models/4384/dreamshaper

[ornate elk]

Thats strange, what's your gpu?

[ornate elk]

Embeddings folder

[silk mantle]

It's not appearing in the textual inversions when I put it there

[light prism]

Since i didnt get any response ill repost my question one last time - any kind of hints are appreciated: Hey i am wondering if there are any ideas to how to modify images with the prompts by describing the operation rather than the output. For example using a image of a robot and a prompt "add an arm" rather then the detailed original prompt extended with the total number of arms "high tech robot, red, photorealistic ..., three arms" and so on. Are there models that are capable or this? In specific cases i got the desired output, but most often the outputs seamed random, especially when asking for "remove XY" it adds the XY and so on

[silk mantle]

Have you tried using inpainting?

[chilly harbor]

rtx 4090

[light prism]

my problem with inpainting is that i only found solutions where you manually have to create masks and i was hoping for a solution where the model can do this alone, so the result is not as human dependant
i am generating multiple versions of images simultanuously and ideally want the ai to merge the results into one on its own in the end
i was even thinking about just training a small model with all images that should be merged and then running this model and generating only one image (which hopefully contains the most commen attributes from all) - but this seams to be overkill for the desired goal

[ornate elk]

Okay then you only need --xformers

[manic storm]

anyone have this problem? The result messing up when I used any lora
This is a depth lora. Better to show how its mess up.

original

Already reinstall sd/reboot computer

[ornate elk]

Hey there is a model and function called Instruct pix2pix that can do this type of stuff. But its not to powerful

[ornate elk]

Try reducing the lora weight

But why using a depth lora when there are Controlnet depth models?

[manic storm]

just used to show you the problem.

lora weight is default 20

I tried 20-30
still not work

[ornate elk]

20 ?

Can you show me the prompt with lora ?

[manic storm]

I used the same setting to gen yesterday. Its ok. But it suddenly mess up

[ornate elk]

What webui do you use?

[manic storm]

all lora mess up

[ornate elk]

Your lora weights is 1.0

Try set it to 0.5 or 0.8

Its in the prompt

[manic storm]

I have already use sd half year. First time have this problem

It is a bug I think. Not prompt or setting problem

I reinstall sd and keep everything default

just simple change the model and lora

https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/7851
I have same problem from this post

[iron hamlet]

anyone knows how to merge 2 sdxl models through commandline no webuis?

[ornate elk]

What's your webui version?
What's your GPU and what's inside your webui-user.bat?

[manic storm]

where to check webui version?

my gpu is 1070 max-q

[silk mantle]

Made another one using the embedding preset and it shows up now, but doesn't seem to be affecting the result at all. Is there some keyword or setting I should be using that I don't know of? Currently clicking the embedding in the GUI before generating

[manic storm]

maybe I gen too much makes my gpu broken?

[ornate elk]

Scroll down to the bottom of the browser

There you see the version numbers

Of torch, xformers and the webui

[ornate elk]

For embeddings you only need to include the embedding name In the prompt

[manic storm]

version: v1.9.3  •  python: 3.10.8  •  torch: 2.0.1+cu118  •  xformers: 0.0.19  •  gradio: 3.41.2  •  checkpoint: be1d90c4ab

[light prism]

thank you i will look into that

[silk mantle]

Hmmm... do you know of a generic reason it might not be doing anything? I trained it on images of me, but the people being generated look nothing like me at all

[ornate elk]

OK your torch and xformers version is a bit old.
You can try to delete the venv folder and relaunch the webui-user.bat.
That should download the latest torch and xformers.

[real willow]

Hi 👋 is anyone using sdwebui API faced this issue on 1.9.4?
Looks like when we use --api --nowebui, it initializes FastAPI, however I get 404 for sdapi/v1/txt2img endpoint.

2024-07-02T13:23:09.211831145Z Startup time: 8.3s (import torch: 3.3s, import gradio: 0.7s, setup paths: 0.9s, initialize shared: 0.2s, other imports: 0.7s, load scripts: 1.7s, lora_script.py: 0.9s). 2024-07-02T13:23:09.247048698Z INFO: Started server process [8] 2024-07-02T13:23:09.247170051Z INFO: Waiting for application startup. 2024-07-02T13:23:09.247527519Z INFO: Application startup complete. 2024-07-02T13:23:09.249085056Z INFO: Uvicorn running on http://0.0.0.0:3000 (Press CTRL+C to quit) 2024-07-02T13:23:09.282696391Z Creating model from config: /stable-diffusion-webui/configs/v1-inference.yaml 2024-07-02T13:23:09.415919820Z INFO: 127.0.0.1:50820 - "GET /sdapi/v1/memory HTTP/1.1" 200 OK 2024-07-02T13:23:10.067623378Z INFO: 127.0.0.1:50822 - "POST /sdapi/v1/txt2img HTTP/1.1" 404 Not Found

[silk mantle]

And giving just the embedding name with nothing else resulted in this

[manic storm]

this one?

[ornate elk]

Yes that one

Do you have python installed and set in path?

[manic storm]

should be installed? because I can run sd normally

I can gen good image yesterday

[ornate elk]

What do you launch normaly?

[manic storm]

just suddenly everything mess up

i use this

[ornate elk]

Ahh this

Can you check for updates?

[manic storm]

already checked

is that possible my gpu are broken?

[stark nymph]

Hello all, I have a 3070 and 3060ti, is there a guide to have the webui use both cards on Linux? A lot of the information I've found on Google is not explicitly helpful.

[ornate elk]

I dont think so.
Can you generate normal images without using a lora?

[ornate elk]

Hey, you would need to use two separate webui instances

[manic storm]

yes. No problem if not using lora

[ornate elk]

Okay, and if you use an other lora ?

[manic storm]

same problem

I tried different lora

[ornate elk]

Or try educe the lora weight by changing the lora:loraname:1 to 0.5 ?

Also you need to use a negative prompt

[manic storm]

But I use lora:1 have no problem before

[ornate elk]

Not every lora will work with :1

[manic storm]

lower the lora will make my result different

I used that those lora for a long time.

[ornate elk]

Some loras are overtrained and need a lower value

[ornate elk]

You have an example of an good image with the same settings?

[manic storm]

wait

[ornate elk]

So I can check the metadata

[manic storm]

I forgot what resolution is this, Just quick gen once in 512px

maybe its steps:30

[ornate elk]

Okay, but for that image you used a negative prompt

That helps getting non corrupt images

Drop the image into the PNG-Info tab and you see the settings you used.

[manic storm]

I know that is different prompt, but those 2 prompt I had used for a long time. I know they are works before.

this one with same prompt and steps

[glad fossil]

i'm wondering if someone can help me. for starters i'm using automatic1111 webui, and when i run it with webui.sh, the ui runs but any models besides pruned v1.5 doesn't show when i refresh the stable diffusion checkpoint. also I install xformers but it doesn't detect it in my system

[wind tusk]

Hello community, I am a beginner currently

studying how to run Stable Diffusion on a Mac M2 8G by Core ML. I successfully managed to run it and attempted to add a module called Cetus Mix for use. However, I always end up with an image that is completely filled with noise, no matter what prompts I use.

Please give me some advice I would be appreciated

ie. Both the stable diffusion model and the Cetus Mix mod all at version 1.5 and converted to core ml (mlpackage/ modelc)

[knotty turret]

guys anyone know how i can use or find the model PonyXL but i want to use it with FP16

its a thing i can change in sd? idk

[karmic crown]

forget core ml and converting models
install https://github.com/AUTOMATIC1111/stable-diffusion-webui
but use my installation script https://github.com/viking1304/a1111-setup
you might need some parameter tweaking after that, because you have 8GB

[pale salmon]

sup, is there a reason why my stabble is very slow ? maybe is my wifi ? or my gpu is dead ? i updated driver too

[cinder ermine]

sorry for the late response, turns out i had to select sdxl in the top right

[ruby yoke]

ah. okay. glad you got it figured out

[cinder ermine]

got another question now, where do i adjust steps?

[ruby yoke]

steps?

[cinder ermine]

sorry meant repeats

[ruby yoke]

here are the docs https://github.com/Nerogar/OneTrainer/blob/master/docs/QuickStartGuide.md you might want to book mark that and refer to it as you go

[cinder ermine]

ty

[ruby yoke]

welcome

[ornate elk]

you need to add --xformers to the webui-user.sh at the line commandline_args=

and outcomment it

[ornate elk]

pony is fp16

[ornate elk]

hey, whats your gpu?

[pale salmon]

GTX 1080 NVIDIA

[fair oxide]

@ornate elk have you tried running sd3 with a1111 after the webui update?

[ornate elk]

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

[ornate elk]

nope

[fair oxide]

there is a reddit post https://www.reddit.com/r/StableDiffusion/comments/1dh0z5v/sd3_support_has_been_added_to_automatic1111/?rdt=40700 but im using comfyui

[pale salmon]

like that ?

[ornate elk]

oh your using forge

[pale salmon]

yeah

[ornate elk]

wouldnt recommend, as its outdated

better use Auto1111

[pale salmon]

https://github.com/viking1304/a1111-setup

this ?

[ornate elk]

This is my guide:
https://github.com/CS1o/Stable-Diffusion-Info/wiki/Installation-Guides#nvidia-automatic1111-webui-stable-diffusion-webui

[pale salmon]

okay will try, ty

[ornate elk]

no problem 🙂 there you need the --xformers --medvram-sdxl --no-half-vae
at the line set Commandline_ARGS=

in the wbeui-user.bat

[fair oxide]

just updated a1111 but something is broken when i go to load sd3 model

[pale salmon]

ok ty man

[ornate elk]

thats old news, it was available since a few days in the sd3 testing branch. and its still not in the stable release

i only will test it when its stable

because right now there are a lot of things that have to be fixed

[fair oxide]

ahhh that explains it why its broken when i tried it now

[pale salmon]

@ornate elki got 3.11 python, is there any prob with this btw ?

[ornate elk]

its not recommended but some Users said it will work with 3.11 too

if not you can downgrade always later

[pale salmon]

okay fine ty

done lets try

[knotty turret]

ah rlly?

lol tried to convert it to fp16

where i can check it

[ornate elk]

if its a 6gb file its already in fp16 format

[knotty turret]

ok

[knotty turret]

btw right now i using webui forge because i cant load sdxl models on a1111. For the lastest update do you think still worth keep webuiforge? or try to make something to use normal a1111 sd

[karmic crown]

that is for macOS only

[ornate elk]

whats your gpu?

[knotty turret]

rtx2060

[ornate elk]

should work in auto1111 too with --xformers --medvram --no-half-vae in the webui-user.bat

also you need 16gb Ram

[pale salmon]

oops okay lol

[knotty turret]

ok but there any advantage using WebUi Forge?

its better back to a1111?

[ornate elk]

nope, since forge will undergoy changes and wont get any updates for normal users, its only advantage is that the current outdated version has a good performance on lower end gpus

[knotty turret]

in your opinion for the rtx2060 forge or a1111

[ornate elk]

i would go with auto1111, but you can also have both installed

[knotty turret]

"also have both installed" yeah good ideia

i gonna do it thx man

[knotty turret]

let mt try this see if work

[knotty turret]

@ornate elk tried to generate image

the model loads fine now but cant generate images for some reason

[ornate elk]

what was the resolution you tried?

[knotty turret]

832 x 1216

[ornate elk]

whats in your webui-user.bat?

and did you launched it with that bat?

[knotty turret]

lauching with that one:

[ornate elk]

can you tray 768x1024 ? as reolsuiotn

[knotty turret]

ok

[ornate elk]

but restart the webui-user.bat before trying

[knotty turret]

Ok

[wild jungle]

Forge is obsolete now??

I did not know this damn

I was gonna reinstall it too lol guess I should just switch back to A1111?

[karmic crown]

https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/801

[ornate elk]

for normal users yes, it will be a test webui in future to test new techniques (that are not cmpatible with extensions and stuff)