🤝｜tech-support

[shadow heron]

so close it and continue with furhter steps?

[main gate]

@ornate elk sorry it's gonna take a sec my breaker blew so i have to regenerate lol old as breaker box

[ornate elk]

oh okay

[main gate]

now im havin a whole new issue before i even get there where img2img super saturates everything and just breaks the image

so i cant even get to inpainting now lmao

[ornate elk]

make sure you whitelist the webui in any browser adblocker

they often break img2img

[main gate]

i turned off all my ad blockers

its like really broken it takes it from this

to this

[ornate elk]

wtf, can you share the settings of inpainting?

[main gate]

thats not even inpainting that's just img2img

it wasn't doing this earlier idk what changed

[ornate elk]

what model and vae do you use?

also does txt2img still work normaly?

[main gate]

abyssorangeremix and i was just told to use grapefruit as the vae i don't really understand how vae works

[main gate]

yes

text 2 img is what generate that first image

ok wait no

text 2 img is broken now too

should i just restart it

[ornate elk]

yep try that

[shadow heron]

So I completed all the steps, how do I know my gpu is being used

[main gate]

ya its just completely broken now

[ornate elk]

if the vram usage of it in the task manager is above 4gb

and the images should take seconds

the first image will take 15-30 mins to initialize zluda, after that its fast

[shadow heron]

getting this error now @ornate elk

I think there is a command to disable thatn o?

[main gate]

ok the sampling method was the problem

[ornate elk]

can you show me the full cmd log?

[main gate]

i changed to dpm and it fixed

[shadow heron]

just copy and paste the cmd log?

[main gate]

actually ogt a pretty dope picture now

[shadow heron]

@ornate elk here https://pastebin.com/giGTtVCY

[main gate]

now to figure out why img 2 img just changes saturation

[ornate elk]

can you do the following:
open up a cmd and type
pip cache purge
then delete the venv folder and relaunch the webui-user.bat
Also have you restarted the PC once after installing hip sdk?

[shadow heron]

I believe so

I'll try the rest

Just open up a random cmd window (not matter what directory cmd is in) and do the pip command?

[ornate elk]

probably due the vae or some other setting, but thats not inpainting now?

[main gate]

once i fix img 2 img then i can try and figure out inpainting again maybe i can show the inpainting problem anyways let me see

ok the vae was the sazturation issue

[shadow heron]

Note I didn't exactly restart the pc after installing hip sdk but rather at this step

[main gate]

so if i highlight like the arm @ornate elk

it just does that

and changes the color

base img

[ornate elk]

whats your GPU?

[main gate]

rx 7900 xt

[ornate elk]

ah okay

[main gate]

20g of vram

[ornate elk]

your on DirectML then

thats the issue

[main gate]

is there a way to fix that

[ornate elk]

yep, reinstall the webui with Zluda instead of DirectML, that fixes everything. and speeds up image genertion by a lot and uses less vram

[main gate]

ummmmm is there a video you could perhaps link on how to do that

[ornate elk]

there is no correct video of the setup, so i made a step bysetp guide here:
https://github.com/CS1o/Stable-Diffusion-Info/wiki/Webui-Installation-Guides
Go with Automatic1111 with Zluda.
Also you need to have the AMD Driver on anything lower than 25.5.1

im on a 7900xtx

@shadow heron hows it going?

[shadow heron]

an error occured while running the user bat file after deletion

I'm redoing the process again

[ornate elk]

ah ok

[shadow heron]

@ornate elk stopping at this part with this errror

[ornate elk]

relaunch

then that error shouldnt appaear again

[shadow heron]

@ornate elk this might take a lil while?

oh its up now

running the first test on the web gui right now. by drawing a simple yellow circle. on txt 2 img. so far its taking a while but thats normal like you said rite?

[ornate elk]

looks good

yea the compiling will take some time

15-30 mins max

[shadow heron]

well dam lol

also are you an admin on this server?

[main gate]

im confused it's failing to install open_clip

[shadow heron]

I'm not sure if this question is appropriate to ask

[main gate]

Getting requirements to build wheel: started
Getting requirements to build wheel: finished with status 'error'

stderr: error: subprocess-exited-with-error

Getting requirements to build wheel did not run successfully

this is where it says error and not done

[ornate elk]

im a community guide of this server

[shadow heron]

@ornate elk also I see 5gbs being used of my gpu memory, that means that my gpu is being used by stable right?

[shadow heron]

oh ok, can I ask you a question via dms on what is and isn't allowed on this server?

and the reddit?

or Is it ok to ask here

[ornate elk]

im not part of the reddit stable team

but you can ask here or in dm

[ornate elk]

can you show me the full cmd log?

[shadow heron]

I'll bring this to the general chat to not clutter this up

[ornate elk]

wrong python version installed

you have 3.13 but you need 3.10.11

[main gate]

can i just install 3.10.11

or do i have to go through and uninstall pythong 3.13 first

[ornate elk]

you have to uninstall 3.13 first

after that install 3.10.11 and then delete the venv folder and relaunch the webui-user.bat

[lime pivot]

Hey guys, I'm looking for someone who can inpaint this character's hand for me. I'm having a hard time and would really appreciate it. 😄 the 2nd image with the black hand is kind of my desired result it doesnt have to be exact. just want the guys hand to touch the floor naturally.

[fierce tulip]

Hey, i've recently switch from a 3090 to a 9070XT due to hardware failure. I want to try Stable Diffusion image generation. Does anyone know a good guide where i can set it up for AMD?

[urban orbit]

I am not sure if it has changed over the years but regretfully Stable diffusion is simply extremely biased towards NVIDIA, I am told you can get significantly better performance gains with AMD GPU if you are runnign it on linux but I have never tried it myself so I can't vouch for it. A1111 and Comfy's own git hub have instruction on how to install them on each OS

[open fox]

Look at the pinned messages in this channel. There is a guide on how to use zluda to get stable diffusion working on amd
https://github.com/CS1o/Stable-Diffusion-Info/wiki/Webui-Installation-Guides

[sleek flicker]

which is installations guide for forge M3 Macbook ?

[tawny saddle]

bet nu one can guess wut my commit hash is uvU

[shadow heron]

any way to make it so that (without setting brave as my default browser), so that when I load the webgui up, it automatically goes to brave instead of firefox?

I use firefox for general basic day to day stuff and prefer to use brave to ai stuff

[urban orbit]

if we are talking about A11, I haven't found one, but what I do instead is that I turnedd off autolaunch in the settings and saved the access port (deafult is 127.0.0.1:7860) as a bookmark on the browser I want to use it on. So affter launching the webui I just open my preffered browser (in my case librewolf) manually and click on the bookmark

[shadow heron]

thats probably a good workaround. I already have it bookmarked on brave

[main gate]

@ornate elk I love you btw took a few hours of troubleshooting but it gens so much faster on zluda

[ornate elk]

Thanks ^^ yea its totally worth to set it up even if its not that easy at first.

[unkempt meadow]

No Not really

[vocal burrow]

Gonna mention it again, just in case.
stability staff have the according discord roles to show for it.
Checkboxes ☑ ✓ ✅, or "staff/dev/support/super hero/etc" in the bio are not a proof of anything.
There is no "external support discord", so don t get scammed in DMs.

[sleek flicker]

Hi need help - please - just started -
which is the best Instructions to use for Macbook?
do i need to install controllnet in Forge ?
Where do i find the settings or description how to change something or creative upscale in old fotos ?

[vocal burrow]

which is the best Instructions to use for Macbook? check pinned messages and look for viking1304's github.
do i need to install controllnet in Forge ? no it should be included
Where do i find the settings or description how to change something or creative upscale in old fotos ? gonna need you to need more precise there.

[open fox]

First of all, do not react to private messages that suggest you to join any kinid of support discord !

It is a scam to get to your account data etc.

I would suggest to look at tools like draw things if you are not used to work with the terminal application and manage python environments. Otherwise you could use comfyui or invoke which work both with mac

[vocal burrow]

yeah I m done with him

hammer time

[sleek flicker]

thanks - i already use draw thinks including img2img - but want to upgrade as inpaint functions are very limited

[open fox]

Yes for inpainting and more creative functions you might have a look at invoke or i use comfyui and the krita ai plugin.

[sleek flicker]

in draw things there a script creative upscale - which quickly changes the quality of any picture i upload and upscale it , best case with face restauration - is there anything like this too ?

in spaces there is a plug in face swap - which i have installed - but when i press launch it just showing the time running but nothing else happens ?

[open fox]

Well with comfyui you could use the detailer or tiled upscale to get a larger and more detailed image. Face swap would be using pulid or other models together with sdxl or flux-

[sleek flicker]

ok lot of thx for your help i will try

[lime pivot]

does anyone know which image i should put as reference image for depth_zoe in controlnet ?

[vocal burrow]

none of those.
depth references images should be a gradient of grey. The whiter the pixel, the closer it is.

eg

[vocal burrow]

preprocessor should be able to create it for you using this one. The grid might throw it off a little.

[lime pivot]

thanks will try it our rn !

@vocal burrow sorry i have another question do you know what the resolution should be? the output is cutoff. usually the resolution is WxH but here its just 1 thing so idk what to put. the purple reference image is 1604x612

[vocal burrow]

change this

[lime pivot]

aha! ty then for the resolution is it just the higher the better ?

[vocal burrow]

I d suggest sticking to the output resolution (before upscaling)

[lime pivot]

thank u !

[frail grove]

why flux fp8 takes so different time to generate an image

[gentle lava]

Hey, can you guys help me set up Stable Diffusion on my RX 580?

[main gate]

does anyone know if it's normal for it to say compilation in progress over and over and take forever when using ADetailer for the first time

[ornate elk]

Can happen when zluda needs to compile stuff for the first time

[ornate elk]

Hey, in the pinned messages first link You'll find the Guides for AMD.
Go for Auto1111 or Forge with Zluda.
The RX580 needs Hip SDK 5.7 to work with zluda and the Guide lists every step.

[gentle lava]

im trying directml

lets see if it will work

[ornate elk]

DirectML will also work and is easier to setup but is slower and uses way more vram than zluda. So your very limited on the resolution and models you can use.

But yea try it out and see if it works. And you can also setup the webui with zluda later too if needed.

[main gate]

ok now im confused control net is installed and everything and i can use it but when i click generate image it just ignores control net completely like it doesn't exist

[azure magnet]

How to fix it?

[harsh moss]

is there a way t use stable diffusion with intel arc a770 16gb?

[ornate elk]

Yes with SDnext

[ornate elk]

You need controlnet models too

[main gate]

Ya i needed a different type of model i fixed it

Now to figure out animatediff

[neon cedar]

I'm getting:
RuntimeError: Input type (float) and bias type (struct c10::Half) should be the same
did git pull and deleted venv and still getting that error

[ornate elk]

Can you show the full cmd log?

[ornate elk]

What's your AMD Adrenalin driver version?

[ornate elk]

25.5.1 and 25.6.1 are currently not supported by Zluda.
Better downgrade to 25.3.1 or 25.4.1

Then it will work again

[neon cedar]

thanks it works now

[lime pivot]

@ornate elk is there a way to optimize forge sdxl? im running on 8gb and my webui crashes a lot with controlnet

[vocal burrow]

use NeverOOM

[lime pivot]

thanks!

[pastel knoll]

Stability.ai how to get free trial cuz wheni try to get it it just ask me to pay

[night locust]

Is anyone able to use ComfyUI-Easy-Use without issues? When I have that one enabled, it breaks the menu that popups when dragging out noodles from nodes. Or that menu works but I cannot access "search" from there, and also when I click away the menu the noodle stays frozen in place and I can't do anything until I press Esc to make the noodle disappear.

[harsh moss]

will it run stable diffusion on my cpu?

[ornate elk]

no if you install it correctly and read this documentation it will use your intel ARC a770:
https://github.com/vladmandic/sdnext/wiki/Intel-ARC

[harsh moss]

thank you

[ornate elk]

no problem

[harsh moss]

i downloaded sdnext and realistic vision v6

but when i try to generate something it says no model selected

how can i load the model?

okay i think i figured it out

[mystic dirge]

has anyone here used fluxgym?

Im having major problems

[sonic nexus]

RuntimeError: CUDA error: no kernel image is available for execution on the device CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
i use 5070 ti

how to fix it

[ornate elk]

If you try to use Auto1111 then you need to follow these steps:
#🤝｜tech-support message

Or install Forge Webui, it works directly with 50 series GPUs

[dense topaz]

Hi i'm trying to generate two separate individuals with different facial expressions but it keeps giving me both of them doing the same one

[ornate elk]

Hey, Make sure adetailer is not enabled. Also try use the regional prompter extension

[solemn atlas]

does someone know what it turns out like this? I didnt changed anything, open webui and this is what i get

comfyui*

[empty cipher]

I know you probably get this all the time, but

Best setup and settings for running on an M3 macbook?

It takes about 45s per 768x768 images for me

Or a suitable SD alternative

[azure magnet]

Help please

[viscid dagger]

hi guys i just got this error while trying to upscale an image for making a poster, maybe you can help fix the issue or suggest a better way of upscaling alltogether. I have an image (pulp fiction poster) that according to windows preferences is 1995x2962 with 96dpi. i want an image that will be 300 dpi at 27"x40". i attempted to run automatic1111 (i downloaded it yesterday and let it self install), i used LDSR at 4.05xxx which was the mulitplier i needed to get 300 dpi, i then recived the following error " RuntimeError: mat1 and mat2 must have the same dtype, but got Half and Float"
not sure where i should post this so ill post in a few catagories feel free to delete as necessary

[loud pond]

Not sure why but regional prompter is acting up for me. Im trying to generate 3 people each with different prompts but for some reason im just getting one person with all of the prompts mashed up together. its weird...
If someone could help that would be greatly appreciated thanks.

[azure magnet]

No

[vocal burrow]

do it I dare you

you fool....

anyways

[vocal burrow]

don't follow their scam link

[azure magnet]

Yes. I see

Try to scan me, he should be fool

[random pulsar]

can someone help me fix my ui

nvm i fixed it

[limpid heron]

I’m trying to make a comfy ui work flow from image to video with anime diff evolve and I can’t figure it out any help would be appreciated

[sudden pike]

@earnest jungle ....

[sudden pike]

try to reduce resolution

that happens to me when i get OOM issues

AI shit takes SO MUCH VRam lmao.
Also if the Drivers are shitty too, the VRam Management can suck ontop so it can use more than it normaly would lol. (AMD problem for me atm)

Also Question to anyone here.

Can Stable Diffusion generate an Image in "sections".
(ill explaion what i mean)
So if you have a trained Lora and it generates a close up. it has quite well details.
But if the Object or Character is "far" away from the "camera" it get so much artifacts.

So my question is. Is it possible to Generate like 4 parts of an Image in Higher Resolution that get put together and scaled down to the set Res?
Like you generate an image in 1024x1024.
It gets split into 4 parts that are 1024x1024 each. Put together as one Image and then scaled down to 1024x1024 in total to get the details in

[ornate elk]

How did you installed it?

[random pulsar]

@ornate elk i already fixed it, it was an issue with one of the file names

[zealous gorge]

Hi

Anyone help me....

🧵 Stable Diffusion + Karras Incident Summary (with GPU info)

1. Installed Python 3.10.6 after issues with newer versions.

2. Set up Stable Diffusion WebUI (v1.10.1) successfully.

3. WebUI worked fine with model (dreamshaper_8.safetensors) and default samplers.

4. Wanted DDIM++ / Karras samplers, so tried adding a custom sampler.py script or extension.

5. Script needed xformers, so I tried installing xformers==0.0.37, but it doesn’t exist for CUDA 11.8.

6. Tried various .whl files from GitHub and VastAI — none worked.

7. My GPU is NVIDIA RTX 3060 12GB (CUDA 11.8) — compatible but still faced issues.

8. Accidentally pasted multi-line shell commands — possibly corrupted environment.

9. Got error: sampler_names[0] IndexError — UI failed to load due to empty sampler list.

10. After multiple failed fixes, ended up formatting my PC to start clean.

[ornate elk]

For me it sounds that you where on an very old webui version. Because DDIM++ and Karras samplers are already available in the webui

Check the first link of the pinned messages for the install guides. Go with Automatic1111 or Forge Webui

[random night]

hmm, i wonder if strix halo chip can use comfyui or forge with zluda

[ornate elk]

Yes it can I think.
But AMD is currently working on "TheRock"
With that strix halo can be used without zluda in Comfyui with the same performance.

[random night]

wait, that's very good news

[ornate elk]

Here is a source:
https://x.com/adyaman/status/1926368074866757857

Here is the repo of "TheRock"
https://github.com/ROCm/TheRock

Aaryaman "Jam" Vasishta (@adyaman)

comfyui running out of the box nicely AMD Strix Halo on windows+ROCm.
10.5 it/s on the default sd1.5 workflow with `python https://t.co/E0xr4ewo8d --use-pytorch-cross-attention`

I've prepared a pytorch wheel with @scottttw (see next tweet for link)

X · May 24, 2025, 8:01 PM

[random night]

it's still on progress

let it take a while then

[ornate elk]

Yep when they provide prebuilt packages for Windows, it should be easy to setup and implement

This could eliminate the need of zluda for ai tasks

[random night]

yeah, hope so

[manic ledge]

Hey guys, this might appear basic as hell. I just started with sd yesterday and here is my situation. I have real images/backgrounds that i took in real life. I also have created some characters/models in stable diffusion and I would like to place them in the very real backgrounds that I took without altering them. How do I go about doing this? I have tried inpainting, imgtoimg with my background but my character never comes out quite right. (I only have faces for the characters) any help would be greatly appreciated. Thanks,

[potent sun]

does anyone here have experience mixing loras?

is there any trick to it, like changing the lora strength?

[fallow knoll]

Bingo

That's exactly it

There's still a limit to how many loras you wanna stack, but as long as you adjust strengths to keep their impact balanced and don't use conflicting loras, you'll be fine.

[loud pond]

from what ive done if i want to use a ton of Lora's ill just reduce the strength to like 0.3 per

[fallow knoll]

Fair warning though, some LoRAs are aggressive and don't play well with others, so it's difficult to get them to not overpower the models base image style and the other loras

[potent sun]

If Im mixing for example an artstyle and a character, I put the style at 1 and the character as 0.5, because there is more marging for error in the character

[fallow knoll]

Most people will say you should only use 1 style lora, 1 or 2 quality loras, and then concept loras as needed.

Or character

[potent sun]

but dunno if there is a sweatspot

assuming both loras work at 1

0.1 definitely is not it

[fallow knoll]

I typically set character loras between 0.7 - 0.9, as i intend for that character to be well represented. I usually will leave a style lora at 0.5-0.7 depending on the model. If your mixing style loras, you should adjust the weights down.

[potent sun]

if they are both trained on the same base model they should work

[fallow knoll]

Style loras are usually more aggressive, since they are intended to affect the entire composition, not just a character or pose

[potent sun]

what I dont know if you can use different loras of different checkpoints, but same core model

[fallow knoll]

Of course it depends on the training data and the quality of the lora as well.

[potent sun]

but there is not a set rule then

[fallow knoll]

Not usually, though it gets muddy with SDXL/PONY/ILLUS and NAI stuff.

[potent sun]

what about something like 1.5sd-sxl

you cant mix them right?

[fallow knoll]

Correct, there's no perfect answer or one size fits all weight combination. It depends on the tools your using.

[potent sun]

mmm I was wondering if there is a way I could mix loras trained on different models

[fallow knoll]

Not usually. A lot of times if you try it won't ruin the photo but the different architecture lora simply won't do anything. Your log/terminal will show it failing, but the image will generate regardless.

[potent sun]

there are many sdxl edits and a lora can have errors if it doesnt have the exact same base model

[potent sun]

can I generate an image with a character lora then pass that same image to the style lora? if both dont have the same base model

Im just not experienced with img2img and inpainting

[fallow knoll]

You could use i2i and different checkpoints yeah, that's a pretty standard workflow for some people that want to style blend

[potent sun]

do you have any tips for img2img and inpainting?

[fallow knoll]

Or you could, if you have comfy, probably build a workflow that automatically switches and loads loras/checkpoints at a set step number. So if your using 30 steps for the image, it would swap to another model to finish say, the last 7 or something.

[potent sun]

I would like to know how to fix generation mistakes but im unable to do so

[fallow knoll]

I can only tell you to look at the readme files for the extension githubs, or YouTube. Explaining img2img and inpainting is a lengthy topic, and it's easier to just test it and learn from your own results.

I use iterative upscaling in latent pixel space to correct my mistakes in images, usually. But that's not a feature of forge/a1111

But I also have a second branch of my workflow that does a hand and face detailer pass, then does a post processing upscale for sharpness.

So I don't use img2img or inpainting much.

[potent sun]

I just cant find a good tutorial

for other topics I do fine, but inpainting is a bit weird

[fallow knoll]

Yeah it's a trial and error kind of process. There's no good tutorial other than the basics, which are easy to understand. Getting it to do exactly what you want is a different story altogether

[silent oak]

any good tutorials for face swaping img2img on SD forge ui?

[slim vine]

Is there anyone who can help me get stable diffusion running with my 9070XT

ive been trying for hours

[ornate elk]

Hey sure. Checkout the first link of the pinned messages for the AMD install guides

[slim vine]

I'm on this guide https://github.com/CS1o/Stable-Diffusion-Info/wiki/Webui-Installation-Guides#amd-automatic1111-with-zluda

I've gotten it to launch

I'm generating my first image and in my CMD it says Compliation is in progress. Please Wait.

[restive sparrow]

A1111, last 2 days my computer has started to hard halt when gerating. I know you can get momentary halts during the peak of a generation but these ones im getting hard halts. And eventually the webui command window eventually just looses its connection and says "press any key to continue"

Im just wondering if this has happened to anyone. cheers

Only notable thing I did recently was updated to May win cumulative update. I had the 24H2 update sitting there so I installed it tonight but that didnt change anything.

[slim vine]

This guide got me working!

[ornate elk]

Perfect 🙂

[ornate elk]

Open up a cmd as admin and run
SFC /scannow
This will fix windows errors that sometimes are caused by updates.

Then make sure your not running Wallpaper engine or Games when using SD.
Also what's your GPU and how much RAM do you have?

[silk hawk]

anyone know how to fix this error? i installed pytorch with cuda 12.8

gpu is rtx 5090 and cpu is amd

[ornate elk]

If you try to use Auto1111 then you need to follow these steps:
#🤝｜tech-support message

Or install Forge Webui, it works directly with 50 series GPUs

[silk hawk]

hm, forge webui is just a UI right? i can still use all lora's and checkpoints that i use with auto1111

i might try that

[ornate elk]

Yes forge is a fork of auto which is a bit more updated and supports flux as models too

[silk hawk]

so what is flux?

[ornate elk]

Flux is a model type like 1.5, sdxl, pony, etc

Its bigger, trained on higher resolution and can produce very good images. But needs a lot more vram.
But with a 5090 its no problem

[silk hawk]

seems like im getting the same error with forge

[ornate elk]

Can you provide the full cmd log?

[silk hawk]

sure,

where are the logs stored

[ornate elk]

Nowhere xD just copy the cmd text ^^

[silk hawk]

it says pytorch doesnt support the 5090, but i went ahead downloaded the latest pytorch with cuda 12.8

[ornate elk]

When do you get that error?

[silk hawk]

when generating an image

or when running hte run.bat

[ornate elk]

Whats the model you try to use ?

[silk hawk]

ponydiffusion

[ornate elk]

V6 ?

[silk hawk]

yup

V6 XL

[ornate elk]

Okay that should work.

Do the Following:
Install python 3.10.11 64bit
No need for uninstalling 3.10.6 just upgrade.
Then open up a cmd and type
Pip cache purge

Then delete the venv folder inside the webui folder.
Then relaunch the run.bat

[silk hawk]

alright, did all that, now when i run it kinda just stalls here

TypeError: argument of type 'bool' is not iterable
Exception in thread Thread-26 (webui_worker):
Traceback (most recent call last):
File "C:\Users\grins\AppData\Local\Programs\Python\Python310\lib\threading.py", line 1016, in _bootstrap_inner
self.run()
File "C:\Users\grins\AppData\Local\Programs\Python\Python310\lib\threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "C:\Users\grins\Desktop\Forge\webui\webui.py", line 101, in webui_worker
app, local_url, share_url = shared.demo.launch(
File "C:\Users\grins\Desktop\Forge\webui\venv\lib\site-packages\gradio\blocks.py", line 2446, in launch
raise ValueError(
ValueError: When localhost is not accessible, a shareable link must be created. Please set share=True or check your proxy settings to allow access to localhost.

last few lines look like that

yeah, same cuda error when running the image gen

[ornate elk]

I guess your GPU driver is updated?

Hmm the log says it used cached stuff again to install pytorch.
Did the pip cache purge deleted anything?

[silk hawk]

yeah

[silk hawk]

it said it deleted like 520 something files

[ornate elk]

Hmm okay

[silk hawk]

ah wait

no there was a driver update

[silk hawk]

nevermind same issue

even with updated drivers

[ornate elk]

Also have you run the update.bat before running the run.bat?

[silk hawk]

yup, it says its all up to date

[ornate elk]

Okay try this:

Go to system/python inside your forge folder
Open cmd prompt in that folder.
Run ```pip uninstall torch torchvision torchaudio -y```
Run ```pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128```

Then relaunch the run.bat

[ornate elk]

Okay then go into the venv folder and run them there

Oh wait

You have to activate the venv before running the commands

[silk hawk]

oh i just ran the second command in the python folder terminal

installed something

[silk hawk]

how do i do that?

[ornate elk]

Run
venv\scripts\activate

And then the other two commands

[silk hawk]

in the python folder?

[silk hawk]

or the venv folder?

[ornate elk]

Venv

[silk hawk]

didnt work

[ornate elk]

You have to run it in a cmd

Not powershell

[silk hawk]

when i right click open terminal

it opens powershell

[silk hawk]

how do i open a cmd in that folder

[ornate elk]

Click into the address bar of the folder and type cmd and hit enter

[potent sun]

can anyone help me here with illustrious models?

https://arxiv.org/pdf/2409.19946

this is their paper

this part talks about data preprocessing but Im getting a bit of doubts with the order

dunno if this is the best place to talk about these kind of topics

[ornate elk]

Oh them just try
Scripts\activate

[silk hawk]

yup something happened

[ornate elk]

Now enter there the other two commands one by one

[silk hawk]

i think i got it working with your help

ill do some testing

[ornate elk]

Alright let me know how it goes

And I wouldn't recommend using Pony v6, its the worst of all pony models.
And illustrious models are better for Anime stuff.

[fluid trout]

would someone be willing to help with a few steps towards training my own Flux Lora locally?

im trying to follow this tutorial but dont understand some stuff https://www.reddit.com/r/StableDiffusion/comments/1f5onyx/tutorial_setup_train_flux1_dev_loras_using/

[daring reef]

My comfyui is coming blank

[daring reef]

but http://localhost:8188/ is working, why?

[brave frigate]

Can someone help me :c It says RuntimeError: mixed dtype (CPU): expect parameter to have scalar type of Float

[vernal sand]

Im trying to train a lora, Im stuck on this step now, anyone has advice for it?

[fluid trout]

could somebody show me how to set the number of epochs in this workflow here? i dont know why its 59 and how to lower it?

[zealous gorge]

Hello i cannot install reactor on extension why......it is removed ? Whats the best alternative

[sudden hazel]

can i use handyolov9c.pt with forge?

[ornate elk]

This one:
https://github.com/Gourieff/sd-webui-reactor-sfw

[restive sparrow]

There is an additional cumul update I downloaded for 24h2 so let's see if that does anything. But yeah I'll do a sfc scan aswell.

[granite bough]

does anyone know how to fix this error RuntimeError: mixed dtype (CPU): expect parameter to have scalar type of Float

[trail mango]

Paste more of error message please

Are you building something custom or running an extension?

Because the error message indicates you have a function that is receiving the wrong data type

[silent oak]

What’s the current best known method for face swap? I’ve been using forge SD. Is going for comfyui worth it? Kinda lazy to learn new stuff

[tawny saddle]

bru..

google gemini's image creation tho.. gracious me

its unhinged when u get ur prompts right..sum dang good art 😳 👀

I'm impressed and surprised

o-o

[zealous gorge]

Hello.....is this similar to vidmage.ai
Vidmage is so simple

How to get vidmage type on forge ui......

[ornate elk]

Forge can't to video stuff. For that you would need to use ComfyUI / SwarmUI

[zealous gorge]

For image only

For image swap only

Vidmage.io is a face swap tool page just for faceswap ....

Can u get similar on forgeui

[ornate elk]

You can install Reactor for faceswap on forge

[zealous gorge]

The reactor is not there its moved i guess or deleted.....

[open fox]

sorry can't say for forge but with comfyui there is a pulid model for face swapping maybe there is a expansion for forge and this model as well

[ornate elk]

I linked you the correct one yesterday

You just need to click on extensions, install from URL and paste it in there and click install.
Then wait a bit and restart.

[static parcel]

how to download stable diff?

[ornate elk]

Check the first link of the pinned messages of this channel.

There are the install Guides

[zealous gorge]

Install and put in where....extension or in model ?

[ornate elk]

In forge click on extensions, install from URL. And then paste there the URL and click install

[daring reef]

when i use the image to video Wan2.1_I2V_14B_FusionX-Q6_K, the video comes out like this, how can i solve it?
but with Wan2.1_T2V_14B_FusionX_VACE-Q6_K.gguf comes out well

[zealous gorge]

Thanks i did but the result is so worst....i tried sampling steps 20-60.....scale 4-8......sampling eular.....then also not good

Iwant to swipe dr nick to captain america

[chilly knot]

👋

Tell me, will I be able to work in comfyui or a1111 with adequate speed on 5060 ti? Will there be any limitations in work?

some write that the work is very limited and the speed is reduced

[ornate elk]

Hey, the card will work good

It has 16gb vram so thats really good for most stuff

[chilly knot]

thx i want buy it

its ok for with pce 4 ?

[ornate elk]

Sure

[chilly knot]

is it relevant that a special guide is needed to install soft for gui for this series?

i saw comfy video tutorial special for 5000 series

[ornate elk]

Mostly you need to install a specific pytorch version for them

Its not to difficult

[chilly knot]

wan video 2.1 generation time on this card?
real <20 min?

[potent sun]

does anyone here use LoRA_Easy_Training_Scripts
I wanna know if its possibole to add samplers

[limber cypress]

Hi guys! Does anyone know how to fix this error?

Red letters: [WinError32] The process cannot access the file because the file is in use by another process

[vocal burrow]

that's the interresting part

basically means "get a stable connection"

[snow breach]

Would anyone happen to know if when running flux in forgeui, does dynamic prompting work? When I prompt something like [maroon | dark blue | white] in the prompt, the same thing comes out in the output file log and the image won't select one of the options. In SDXL it would simply come out as a single option in the output file log.

[zealous gorge]

Hi ,i just instal forgeui....is there any onkine training course from begineer ....i want to learn frim beginning....dont know where to start from sd1.5 or sdxl or flux...there are 3 option.....and many other like controlnet,vae,lora etc......is there any online compmete beginner course ?

[zealous gorge]

Hello how to install ipadapter extension....i am having problem with it......and also ip adapter for sd1.5 and sdxl..plz anyone...

[pale bobcat]

hello, can someone help me with this error: ModelPatchTorchSettings
Failed to set fp16 accumulation, this requires pytorch 2.7.0 nightly currently

thanks

[ornate elk]

hey, where do you get this and when?

whats your gpu and which webui?

[pale bobcat]

i am trying to use this workflow from https://www.patreon.com/posts/131089902?utm_source=youtube&utm_medium=video&utm_campaign=20250610

my gpu is an RTX 4090 and the webui is comfyui
i am running PyTorch, version 2.8.0.dev20250614+cu128.

[potent merlin]

installed forge but stil cant run onn 5070

[ornate elk]

try this:
Go into webui_forge_cu121_torch231\webui folder
Open cmd prompt in that folder by clicking into the adress bar and type cmd and hit enter.
Then copy and paste these three commands one by one:
venv\scripts\activate
pip uninstall torch torchvision torchaudio -y
pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

Then relaunch the run.bat

[gentle forge]

I am also trying to install on a 5070 and I keep getting an "Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check" but I can't find the COMMANDLINE_ARGS

[ornate elk]

hm then you cant use it or you have to downgrade to pytorch 2.7.0

[ornate elk]

follow my steps from above

also make sure the gpu driver is updated

[gentle forge]

Mine doesn't have forge, it's just stable diffusion.

I have a "webui" folder

[ornate elk]

then same thing but do the commands in the stable-diffusion-webui folder instead of webui

[ornate elk]

whats the folder called where the venv is located?

[ornate elk]

can you reinstall the Webui by following the instructions from here?
https://github.com/CS1o/Stable-Diffusion-Info/wiki/Webui-Installation-Guides
Follow the Auto1111 or Forge guide for nvidia

then launch it once and let it install its stuff, and then you can use the commands to switch to the correct torch version

[gentle forge]

"git" is not recognized

[pale bobcat]

i already tried to downgrade, didnt change anything

[ornate elk]

ypu have to install Git and python

check the Preparation part

[gentle forge]

I got git, thanks

I installed python from the link they provided but I still get this

hold on, might have got it

[potent merlin]

didnt even have venv somehow

what do i need to enter to create?

[ornate elk]

you have to launch the run.bat

[ornate elk]

you need to install python with "add python to path" checked

[vital raptor]

why am i having this error? thanks

[gentle forge]

Yeah, I missed that. I got it though, thanks

[ornate elk]

idk but what are you even using or trying?

it doesnt look good when random stuff is inside Windows/System32

[potent merlin]

after runing run.bat i still dont have venv

@echo off

call environment.bat

cd %~dp0webui
call webui-user.bat
thats whats in

[ornate elk]

try edit the environment.bat file and set create_venv to 1

then relaunch

[potent merlin]

with 1 =

like that?

[ornate elk]

it should look different

can you show me the full environment.bat content?

[potent merlin]

@echo off

set DIR=%~dp0system

set PATH=%DIR%\git\bin;%DIR%\python;%DIR%\python\Scripts;%PATH%
set PY_LIBS=%DIR%\python\Scripts\Lib;%DIR%\python\Scripts\Lib\site-packages
set PY_PIP=%DIR%\python\Scripts
set SKIP_VENV=1
set PIP_INSTALLER_LOCATION=%DIR%\python\get-pip.py
set TRANSFORMERS_CACHE=%DIR%\transformers-cache

[ornate elk]

ah there just set skip_venv to 0

[potent merlin]

why is that on default 1

works finally now

thank you

[ornate elk]

perfect np ^^

[gentle forge]

I am getting an error when trying to train a Dreambooth model that I am missing xformers. How do I install this?

I downloaded xformers-0.0.30 but I do not know where to install it

[ornate elk]

dont use dreambooth its old and outdated

it also breaks the webui

Use OneTrainer or Khoya_ss for Lora or model training. They are both standalone Tools

[gentle forge]

Okay, I'll try OneTrainer

thanks

[granite bough]

RuntimeError: Input type (float) and bias type (struct c10::Half) should be the same stable diffusion

does anyone know how to fix this error i've already reinstalled sd twice

[mystic dirge]

Is there any way to use OpenPose with Flux in WebUi???

The .pth model in the Integrated ControlNet is marked for StableDiffusion not Flux

[mystic dirge]

This is so unbelievably fake it isnt even funny

[ornate elk]

What's your GPU and share the full cmd log

[granite bough]

Rx 6950

[ornate elk]

Are you on auto1111 or forge with zluda?

[granite bough]

auto1111

[ornate elk]

Ah okay what's your AMD Adrenalin Driver Version?

25.5.1 and 25.6.1 cause issues like your error

[granite bough]

25.6.1

Ah

Which one should I downgrade to

[ornate elk]

25.4.1 or if you play VR Games then 25.3.1

[granite bough]

Ok

[inner dawn]

Hi, how in stable diffusion forge, inpaint / image to image, to switch from drawing mask to image, to deleting (erasing) the mask?

[vital raptor]

any help will be appreciated

[ornate elk]

You didnt answered my question about it

[vital raptor]

sorry ive missed it, im on win 11, trying to run text to video, ive put it outside system32 now

[rocky dragon]

Is it possible/legal to host a service in a cloud computr instance based on A111K functionality?

I assume the webUI would not be used and a different setup would be required?

Are there guides for this?

[main nest]

Can someone help me install inswapper_128.onnx for FaceFusion 3.1.2 😄

I've been trying for the past hour

[ornate elk]

Which tool exactly?

[ornate elk]

inswapper is already included in facefusion

i updated to 3.2.0 and ist still there

[main nest]

Yea but inseapper_128.onnx isn’t

Ive been searching for that onnx version

[ornate elk]

this is how it looks inside facefusion models folder:

are you on AMD or nvidia gpu?

[main nest]

Nvidia

[ornate elk]

and your missing only the inswapper file?

that one from my screenshot?

[main nest]

Yea

I’ll send you a screenshot

One sec

[honest harness]

Thats me but from the PC

I have them but I don't have option to use it from the FF

oh, its the same model but with two files...

btw when I click on Generate, it starts a loading to 13sec and then stops without showing any preview or result

[vital raptor]

stability ai

[honest harness]

by that I mean @main nest is my second account, I just work from 2 PC's 😄

[ornate elk]

It is there yep

[honest harness]

pass

[ornate elk]

Ah okay can't help with that sry but if you want to use text to video stuff you better setup ComfyUI and there use Wan 2.1 or LTX video

[ornate elk]

Make sure FaceFusion is whitelisted in any browser adblocker

[honest harness]

oh and one more thing, how can I make this to GPU not CPU

[ornate elk]

What's your GPU model?

You have to reinstall it correctly. I can give you my setup bat that will install cuda pytorch with it

[honest harness]

That's my configuration

[ornate elk]

Thats a beast of PC

Do you already have Git and Python 3.10.11 or Python 3.11.9 installed?

[honest harness]

Thanks , and I aim for max quality and realism with all generations 😄

[honest harness]

yup

[vital raptor]

thanks, any tutorial or good guide?

[ornate elk]

Setup guides for comfyui are in the first link of the pinned messages in this channel.
For the video usage and setup guides you should checkout some recent yt videos

[honest harness]

I will be greatfull, but I preffer 3.1.2 FF as I read it has some bugs

[ornate elk]

You can make a copy of the FaceFusion folder and then put in my setup bat then your still on 3.1.2

When I'm at the PC I upload it

[honest harness]

okay, thanks for that

and btw do you have your own setup on how to train a model

[ornate elk]

You can train a model or lora with OneTrainer or Kohya_ss

[vital raptor]

ive installed comfy ui and get this message, there is nothing inside checkpoint folder, where do i get the correct file?

[fallow knoll]

You need to move your existing or download new checkpoints(models) files and put them in the correct folder in the comfyui folder.

Simply downloading comfy doesn't populate it with models/checkpoints

[ornate elk]

Hey, here is the file, put that into the copied facefusion folder and launch it (not as admin).

[fading dagger]

Photography

[honest harness]

still

nothing

[ornate elk]

then try git clone facefusion into a new folder and then run the bat

maybe your version is just broken

[ornate elk]

you need the actual url to facefusion github

git clone https://github.com/facefusion/facefusion.git

[honest harness]

and I still have only CPU

and no generation

[honest harness]

oh and when I add a LoRA to my folder, I can't see them in my extension tab

I have only one visible LoRA

[celest delta]

i'm using wan 2.1, how do i save a video? All i have is this and there's no option to save at all, and the output folder only has a single image instead of an animation

[hardy granite]

Hi guys, I am using Stable Diffusion with AUTOMATIC1111 and I have a question. I have an image with a character and I want that character to appear in all the generated images. I have read about forming a LoRA, but first I need the images in different scenarios, different clothes, different postures... Thats what i dont know how to do. Anyone know how to help me? Thank you to all.

[sand marlin]

Someone maybe trained a wan2.1 t2v lora here and can share a bit of expirance please? My training running but i get just noise in the end results.

[honest harness]

Hello, any idea for this ?

[potent sun]

everryone is saying to inpaint with kirita and invoke, what are those

they arent local, arent they

[ornate elk]

both are local tools

InvokeAI: https://github.com/invoke-ai/InvokeAI
KritaAI: https://github.com/Acly/krita-ai-diffusion

[potent sun]

But they are models, UIs or what

If they are models can't I just use them in forge ?@ornate elk

[vocal burrow]

not sure what your question is.

[honest harness]

Hello, can anyone help me with FaceFusion

Please

[honest harness]

I rebuild FF with git clone command in a new folder and when I tried to FaceSwap nothing happens again, and I don't have GPU option

[honest harness]

just found the problem, its not faceswaping when im using nsfw pictures as target, for which I have no idea why, I do remember it was allowed to FS Nsfw content.

And I still have only CPU as option for EXECUTION PROVIDER

Scam @echo hill

[vital raptor]

Hi, i'm tying to use comfyui in windows with wan2.1 and i get this error about sageattention, how can i get this file? thanks for your help

[main nest]

Is there any support online, Ive been having issues for the last couple of days and still can't fix it

[quaint flame]

how do you ensure consistency in NSFW content when working with Stable Diffusion? i’ve been using savro to adjust body features keen to learn how you maintain realism across multiple generations

[vital raptor]

im using Wan14Bi2vFusioniX model with 3090, all drivers updated, after 67% server exits with no errors, nothing in log. I really dont know what to look for. any help will be appreciated.

[lime pivot]

hello guys, does anyone know if there are significant differences between a 32gb vs 24b vram in image generation? maybe in video/gif generation too

[low escarp]

You can run larger models

Which means more detail

[lime pivot]

@low escarp thanks!

[low escarp]

No problem im kind of researching the same

[winged siren]

I downloaded a detection model for adetailer but I have no idea where to put it so I can actually use it

[ornate elk]

It goes in models/adetailer

[brittle lantern]

Question on the building topic of pcs I’ve some people are picking up the 14900k or KS mainly the KS even tho it’s £600+ GBP primarily as a AI build or a hybrid gaming/AI build is there a particular reason I would assume for Intel u would go with Core ultra series as for the AI only build since it’s mid for gaming but good for AI stuff (I have no idea where to ask this specific question so I’ll ask here since yk AI stuff)

[hallow latch]

This is not an big issue but when I am generating my picture or upscaling I don't get the dimension saved when I hover it. So I guess it's missing the "metadata tags" How do I fix that?

[dapper timber]

Yo

ERROR: Could not install packages due to an OSError: [WinError 32] The process cannot access the file because it is being used by another process: 'C:\Users\dhruv\AppData\Local\Temp\pip-unpack-wkwl4mi6\torch-2.1.2+cu121-cp310-cp310-win_amd64.whl'
Check the permissions.

Traceback (most recent call last):
File "C:\AI\stable-diffusion-webui\launch.py", line 48, in <module>
main()
File "C:\AI\stable-diffusion-webui\launch.py", line 39, in main
prepare_environment()
File "C:\AI\stable-diffusion-webui\modules\launch_utils.py", line 381, in prepare_environment
run(f'"{python}" -m {torch_command}', "Installing torch and torchvision", "Couldn't install torch", live=True)
File "C:\AI\stable-diffusion-webui\modules\launch_utils.py", line 116, in run
raise RuntimeError("\n".join(error_bits))
RuntimeError: Couldn't install torch.
Command: "C:\AI\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 . . .

I am getting this error again and again

[ornate elk]

What's your GPU and python version?

[dapper timber]

3.10.6 and rtx 2050

update: just updated to 3.10.11

[rigid nymph]

im getting an error while trying to install reactor, anyone trynna help me?

[ornate elk]

Alright then open up a cmd and run
Pip cache purge
Then delete the venv folder and relaunch the webui-user.bat

[ornate elk]

You need to install this one:
https://github.com/Gourieff/sd-webui-reactor-sfw

[rigid nymph]

what do i do cause i still get the error

[ornate elk]

Can you show the errror?

[rigid nymph]

@ornate elk

[ornate elk]

Can you show the full cmd log?

[dapper timber]

yo idk if this comes under tech support per se, but I'm having trouble with models taking too much time, i have rtx 2050, should i switch any models? rn I'm using cyber realistic pony

[rigid nymph]

@ornate elk

[ornate elk]

Sry I'm AFK, will check and answer later

[rigid nymph]

aight

[rigid nymph]

@ornate elk you back?

[ornate elk]

Yep right now

Ah your missing the insightface file

Here are the steps to install it:
https://github.com/Gourieff/sd-webui-reactor-sfw?tab=readme-ov-file#viii-for-windows-users-if-you-still-cannot-build-insightface-for-some-reasons-or-just-dont-want-to-install-visual-studio-or-vs-c-build-tools---do-the-following

[rigid nymph]

okay i got it what do i do with the file

@ornate elk

[ornate elk]

the steps that are explained there

oh sry wrong link

this one is the correct: https://github.com/Gourieff/sd-webui-reactor-sfw?tab=readme-ov-file#insightfacebuild

[rigid nymph]

which one is it explained in

?

[ornate elk]

ok delete that stuff again ill explain step by step

do you have the stable-diffusion-webui folder?

if so go into stable-diffusion-webui/extensions folder

[rigid nymph]

this?

[ornate elk]

ah thats the portable one

you can use that but i recommend installing the non portable version

[rigid nymph]

could you send a link to that version

[ornate elk]

but if your using always this one then no problem

[rigid nymph]

oh ok

[ornate elk]

then go into the webui/extensions folder

[rigid nymph]

did it

[ornate elk]

then click into the explorer bar. and type cmd and hit enter

then you should have a cmd that says webui/extensions

[rigid nymph]

yea i do

[ornate elk]

ok then copy and paste this command:
git clone https://github.com/Gourieff/sd-webui-reactor-sfw.git

[honest harness]

Hello CS1o

[ornate elk]

hey

[honest harness]

Do you have time helping m,e with the FaceFusion problem

[rigid nymph]

'git' is not recognized as an internal or external command,

[ornate elk]

oh okay right because your on the portable version

that scammer will be banned

[ornate elk]

you can install Git 64bit from here: (not the portable one)
https://git-scm.com/downloads/win

[rigid nymph]

which one

[ornate elk]

you can leave everything default

its just a tool to download stuff from github

[rigid nymph]

alr i got it

[honest harness]

As before, when I open FaceFusion and insert picture as source of a face and as a target a nsfw body, its not generating anything, not even showing me a preview, BUT when I retry the same action but with non nsfw target photo I get a generation, as long as I know FaceFusion is scaning only the faces and doesn't care about nsfw or not, so why this is a problem?

Second problem which we can fix later is that I tried with your batch file for FaceFusion, reinstalled the FF with Git clone command and still have only CPU option as EXECUTION PROVIDERS

[ornate elk]

now you can run the command again in the same cmd

[ornate elk]

FaceFusion doesnt allow NFSW stuff because they would get banned from Github like Reactor did at first

The other issue is interesting, i will try to get it working on my older nvidia pc

@rigid nymph this is the insightface file you need next: (direct link)
https://github.com/Gourieff/Assets/raw/main/Insightface/insightface-0.7.3-cp310-cp310-win_amd64.whl

Download that and put it into the webui folder

[honest harness]

is there any FaceSwapping extension which allows nsfw or any other option, I remember back then when I was using it in 2022 or 2023 i don't quite remember, I didn't had a problem with it

Hmmm can I faceswap with the batch option and use nsfw or any type of nsfw in FF will not generate anything

[ornate elk]

got the scammer xD

[ornate elk]

there was roop-unleashed which got also removed from github

[honest harness]

meaning, I can't use it ?

[ornate elk]

you can but you need to get the files

[rigid nymph]

here?

[ornate elk]

okay perfect.
Now in the webui folder where you are. Click in the explorer bar and type cmd and hit enter

then copy and paste the following commands one after one:
venv\Scripts\activate.bat
python -m pip install -U pip
pip install insightface-0.7.3-cp310-cp310-win_amd64.whl

[rigid nymph]

ERROR: insightface-0.7.3-cp310-cp310-win_amd64.whl is not a supported wheel on this platform.

it pops up after pasting the 3rd one

@ornate elk

[ornate elk]

hmm okay

then go one folder back and try run the update.bat

then try again

also screenshot if you get the error again

[rigid nymph]

in the webui folder?

@ornate elk

[ornate elk]

hm thats strange

[honest harness]

what if I download Cuda Toolkit, maybe this way I will have GPU in FF

[ornate elk]

not needed normaly

[honest harness]

I remember I had to download it for something in FF before 2 years, but not sure for what

[ornate elk]

i would reinstall the webui the recommended way then also not on desktop due permission errors that can occur.
On the first link of the pinned messages you find the Nvidia Automatic1111 Guide. Try install that and also upgrade to Python 3.10.11 64bit

[honest harness]

Btw does InsightFaceSwap Extension in auto1111 work with Nsfw, in my research I found it does allow nsfw for now

[ornate elk]

yes it works but only if you install the nsfw reactor extension (which also got removed from github) but i have a copy

[honest harness]

Is there a way you can share it with me

[rigid nymph]

is there suposed to be only 1 folder after the download

?

[ornate elk]

yep

[honest harness]

I guess in DM's ?

[rigid nymph]

@ornate elk

[ornate elk]

you have the wrong python version installed

uninstall python 3.13

then install 3.10.11 64bit from here:
https://www.python.org/downloads/release/python-31011/

[rigid nymph]

where do i uninstall it?

oh i got it

[rigid nymph]

k i got it

[ornate elk]

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

[rigid nymph]

bro i deleted it

ion know did i delete it the right way tho

[ornate elk]

i hope you "uninstalled" it and not just deletet its folder

[rigid nymph]

ion know

[ornate elk]

delete the venv folder

then relaunch

if it still says python 3.13 then you need to reinstall python 3.13.5 and then correctly uninstall it again

[rigid nymph]

i dont know did i uninstall it right

[ornate elk]

me neither so maybe reinstall and then use system control pannel to uninstall it

[rigid nymph]

alr

do i press uninstall?

[ornate elk]

yes

[rigid nymph]

alr i did it

[ornate elk]

open up a cmd and type
python --version

[rigid nymph]

python not found

[ornate elk]

okay then now install python 3.10.11 64bit

also check "add python to path"

[rigid nymph]

it is installed

[ornate elk]

click modify

then next

and then "add python to environment variables"

[rigid nymph]

which

[ornate elk]

rerun it as admin

[ornate elk]

can be ignored

[rigid nymph]

i did

evrthing is the same

[rigid nymph]

no wait i got it

alr i added it

[ornate elk]

now rerun the webui-user.bat

[rigid nymph]

alr something is downloading

[flat yarrow]

Hello I wonder if you can help, I've just recently upgraded my GPU from a 4060 to the 5090, and I'm using a workflow that uses Load Flux ControlNet node. A quick Google search revealed that there's being issues with that node, it got me as I though it was a pathing issue, is there a work around for the 50 series for this node.

[ornate elk]

That guy was a support scammer

[flat yarrow]

@white forge I'm not that stupid, if you can't help here then I'm not interested as it helps others too with the same issue, thankyou for your response.

@ornate elk I know, I was very polite though, I hate SCAMMER'S

[ornate elk]

Can't help with Comfyui and custom nodes but you can ask also in the #🧣｜comfy-ui channel

[rigid nymph]

@ornate elk i downloaded the files

[flat yarrow]

@ornate elk Thanx! I have the same question in a few servers, the node is unclaimed in the listing in manager, if it's a script/coding issue it will be upto the author to fix it for the RTX 50 series, and if they will supply an update. But the 50 series is some what in the wild at the moment so I expect issues will be fixed in time. Still doesn't hurt to ask the question though.

[ornate elk]

Okay now go into the stable-diffusion-webui/extensions folder and open up a cmd there by clicking in the explorer bar and type cmd

then copy and paste this command:
git clone https://github.com/Gourieff/sd-webui-reactor-sfw.git

[rigid nymph]

i already pasted it in there before

@ornate elk

[ornate elk]

Ah okay

Then you need to download the insightface face file again or copy it over from.the other folder

[rigid nymph]

this?

[ornate elk]

In the stable-diffusion-webui folder Click in the explorer bar and type cmd and hit enter.
then copy and paste the following commands one after one:
venv\Scripts\activate.bat
python -m pip install -U pip
pip install insightface-0.7.3-cp310-cp310-win_amd64.whl

[rigid nymph]

@ornate elk

it popped up after i pasted it again

[ornate elk]

Okay looks normal

You can try now it reactor Faceswap works

For example in the extras tab If you habe two real photos.
Or in txt2img if you want to generate a person with a specific face

[rigid nymph]

(from scikit-learn>=1.3.2->albumentations==1.4.3) (1.5.1)
Requirement already satisfied: threadpoolctl>=3.1.0 in c:\users\huber\desktop\cai\stable-diffusion-webui\venv\lib\site-packages (from scikit-learn>=1.3.2->albumentations==1.4.3) (3.6.0)
Downloading albumentations-1.4.3-py3-none-any.whl (137 kB)
Installing collected packages: albumentations
Attempting uninstall: albumentations
Found existing installation: albumentations 2.0.8
Uninstalling albumentations-2.0.8:
Successfully uninstalled albumentations-2.0.8
Successfully installed albumentations-1.4.3
CUDA 12.1
Error: The 'onnxruntime-gpu' distribution was not found and is required by the application
Error: (albumentations 2.0.8 (c:\users\huber\desktop\cai\stable-diffusion-webui\venv\lib\site-packages), Requirement.parse('albumentations==1.4.3'))

    +---------------------------------+
    --- PLEASE, RESTART the Server! ---
    +---------------------------------+

Launching Web UI with arguments: --xformers --medvram
C:\Users\huber\Desktop\CAI\stable-diffusion-webui\venv\lib\site-packages\timm\models\layers_init_.py:48: FutureWarning: Importing from timm.models.layers is deprecated, please import via timm.layers
warnings.warn(f"Importing from {name} is deprecated, please import via timm.layers", FutureWarning)
Downloading vit-base-nsfw-detector, please wait...

[ReActor] Downloading config.json to C:\Users\huber\Desktop\CAI\stable-diffusion-webui\models\nsfw_detector\vit-base-nsf[ReActor] Downloading config.json to C:\Users\huber\Desktop\CAI\stable-diffusion-webui\models\nsfw_detector\vit-base-nsfw-detector\config.json: 16.0kB [00:00, 61.3kB/s]
[ReActor] Downloading confusion_matrix.png to C:\Users\huber\Desktop\CAI\stable-diffusion-webui\models\nsfw_detector\vit-base-nsfw-detector\confusion_matrix.png: 48.0kB [00:00, 153kB/s]
[ReActor] Downloading model.safetensors to C:\Users\huber\Desktop\CAI\stable-diffusion-webui\models\nsfw_detector\vit-b

oh okay it works now

but i have 1 more question

is realisticVisionV60B1_v51VAE.safetensors the newest version?

[ornate elk]

For an 1.5 model thats its latest version yes

[rigid nymph]

are there any more realistic versions or is this the best?

[ornate elk]

How much GPU vram do you have?
There are sdxl models which are better but also larger

[ornate elk]

Ah okay then you dont need --medvram in the webui-user.bat

And then you should checkout some realistic sdxl models

For example this:
https://civitai.com/models/312530?modelVersionId=1910174

[rigid nymph]

thanks a lot

[ornate elk]

No problem

Sdxl models are trained on 1024x1024 so dont forget to increase the resolution

[fair oxide]

hello. I'm following these instructions: https://github.com/CS1o/Stable-Diffusion-Info/wiki/Webui-Installation-Guides#nvidia-automatic1111-webui-stable-diffusion-webui

When we have to edit the webui-user.bat

Does it come out like this?

@echo off

set PYTHON=
set GIT=
set VENV_DIR=
set COMMANDLINE_ARGS= --xformers --no-half-vae
Check Task Manager ->Performance ->GPU ->Dedicated GPU-Memory --medvram

call webui.bat
?

[pale ibex]

Any idea why roop isnt being triggered? I've enabled it and I'm using it along with control net for swapping. Even the terminal roop isn't getting logged. I used to be able to do it seamlessly it just stopped working out of nowhere and it's not even giving me any error, just not being triggered at all.

Are there any alternatives to roop? (that is publically available)

[ornate elk]

No just add:
--xformers --no-half-vae

And if you have less than 8gb of vram then add --medvram there too

[ornate elk]

There is reactor for Faceswap as extension for Auto1111/forge etc.
https://github.com/Gourieff/sd-webui-reactor-sfw
Or FaceFusion as standalone tool:
https://github.com/facefusion/facefusion

[fair oxide]

Did it, but it won't open the browser at the end

In fact, when I run it, seems to do nothing at all

[ornate elk]

Can you show me the cmd log?

Or does no cmd opens up?

[fair oxide]

yeah

Sorry, my Windows is brazilian pt.

'“C:\AI\stable-diffusion-webui\venv\Scripts\activate.bat”' is not recognized as an internal
or external command, operable program or batch file.

Says it in the first line

The probem being in the folder I can't see even this folder venv

[vocal burrow]

check if your antivirus blocked the activate.bat script

if so, whitelist it

wait there s not even the venv folder ?

[ornate elk]

Do you have an third party antivirus installed?

[vocal burrow]

(daily reminder to not click sketchy af and non existant "official discord support")

[fair oxide]

No, never had. I believe only the Windows native crap.

[vocal burrow]

what s in your c:\ai\stable-diffusion-webui folder ? screenshot it

[fair oxide]

this

[vocal burrow]

so there is a venv folder....

[fair oxide]

oh yeah

sorry

[vocal burrow]

check if the activate.bat is in there.

in venv/scripts/activate.bat

screenshot what s in there too

[fair oxide]

no, it's not

[vocal burrow]

that looks more and more like an anti virus issue.

rename that venv folder to venvBack or whatever and run webui-user.bat.

Then copy paste the terminal output in here.

[fair oxide]

so far

[vocal burrow]

let it do its stuff

[fair oxide]

still doing

[vocal burrow]

screenshot it just to be sure

@fair oxide

[fair oxide]

sorry, man. I've to split attentions here. My kid is sick.

Now, it seems, it has opened the browser tab

[pale ibex]

i tried facefusion and it works well for now. Thanks

[vocal burrow]

np we all have stuff to do. Welp it's working now like you noticed

[fair oxide]

I just promted something here to test

If you don't mind something more... In this web interface can I work with different LoRAs and Checkpoints?

[vocal burrow]

yup

[fair oxide]

How would that be done?

[ornate elk]

You can go to Civitai.com
There you find models and loras

Use the filter on the right of the models tab

[vocal burrow]

you put them in stable-diffusion-webui\models\Lora and
stable-diffusion-webui\models\Stable-diffusion

[fair oxide]

thanks, I'll try it!

[fair oxide]

now this happened

[vocal burrow]

what settings are you using ?

screenshot your whole webui

[fair oxide]

that

[vocal burrow]

not gonna explain your error but you're using the default sd1.5 model.
1/ it's ANCIENT;
2/ it's trained for 512x512 (more or less 25%)
3/ what gpu are you using ?
4/ try using forge or reforge instead of auto1111.

[fair oxide]

how can I do it?

[vocal burrow]

do what ? use another model ? like cs1o pointed out get one from civitai, huggingface, etc and put it in stable-diffusion-webui\models\Stable-diffusion and select it at the top left of your webui.

[fair oxide]

NVidia RTX 4050 Laptop

6GB

[vocal burrow]

yeah try forge/reforge then. please cf pinned guide for install instructions

[fair oxide]

[NVIDIA] Forge-Webui << this instructions?

[vocal burrow]

yes

[fair oxide]

ok, downloading it

[fair oxide]

what is better? Forge or Reforge?

meanwhile I restarted it, and tryed out...

[fair oxide]

after downloading webui_forge_cu121_torch231.7z, I unzip it anywhere?

[vocal burrow]

forge is more used, you'll get more resources to learn it. But for the end users they're basically the same. Also reforge is on hold I think for now.

[fair oxide]

I ran update.bat and seems not to have added anything into the inferface

[fair oxide]

I downloaded Forge and installed it.

But now it says I don't have any models...

[vocal burrow]

well did you put any in the models/stable-diffusion folder ?

[fair oxide]

I didn't know it would come empty...

Trying to figure out where to download it

[ornate elk]

on civitai.com

[fair oxide]

Mind to share links?

[ornate elk]

here:
https://civitai.com/models

try this for example:
https://civitai.com/models/46422/juggernaut

[fair oxide]

Thank you

[fair oxide]

Do I need to download anything to this: VAE / Text Encoder?

[ornate elk]

No. But for the sdxl VAE you can download this one:
https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl.vae.safetensors
Put it into the models/VAE folder.

[lime pivot]

does anyone know how those video generations are made? ive only used stable diffusion but would like to try it one day. im looking to upgrade my pc rn so i wanna know what specs is needed for those video generations. is it just vram like SD ?

[ornate elk]

Yes vram is needed. The more the better

16gb minimum

[lime pivot]

@ornate elk are other specs like ram (ddr sticks) and cpu important also ?

[ornate elk]

CPU not that much but if you go for a high end GPU then getting a good CPU with 8 cores or more is also good.
But thats just to not bottleneck the GPU in games.
For RAM its important to start with 32gb and if you go for a 24 or 32gb GPU then go for 64gb RAM and always dual channel.

[lime pivot]

@ornate elk thanks for the indepth advice !

[ornate elk]

No problem 🙂
For CPU I would recommend going for an AMD CPU with an AM5 Mainboard (preferable from MSI or Asus).
And the AMD CPU can be something like a 9800X3D if you go for highend stuff.
Or something like a 7600X if you go for budget.
AMD CPUs like RAM that has 6000Mhz and CL30-32 Timings

[untold flame]

What the actual hell is going on with comfy UI

I was in the middle of genning images

it started rebooting endlessly (brand new install from the ground up 2 days ago)

it has now installed 11 node packages I did not want, as well as over 100GB in models I don't want!

I have restarted and stopped it several times, but every time I do, it just keeps downloading other things!

it downloaded all of these, and its STILL GOING

it just downloaded FP16 AND Fp8 Flux Dev, now its installing dreamshaper 2.1

I can't get it to stop!

Please, if anybody can help

@old sky I'm sorry to ping you out of nowhere, but ComfyUI is filling up my whole PC and I don't know why!

Its now installed over 130GB of useless models

it even installed llava v1.5 7b for no reason!

it installed Comfy inpaint nodes, api tools, extra api, magicquill fixed, brushnet, depthanythingv2, reactor, wd14 tagger, workspace manager, and was ns

I just deleted them all and now its installing them all again

I just fully restarted my PC. I HOPE it fixes it

Ok, it started installing everything again. I just went in and deleted everything, and it seems like it installed something called reactor? And thats what was downloading everything I think?

[gloomy wave]

What everyones general generating time for text to img?

I have Nvidia 5070 ti, Comfyui, SD3.5, example workflow and uses the starter prompt (a bottle).
It takes about 89-100 seconds to generate, is this not a little slow?

[azure magnet]

Hi guys, i need an help, i need to install comfyui-manager, how should i do?

[potent sun]

is it true that sdxl is limited to 75 tokens?
how do you train a model with more than 75 tokens in tags?

[ornate elk]

hey, how did you installed comfyui?

[azure magnet]

Fixed, but i have another problem, how upgrade comfyui portable?

[ornate elk]

you can do it in the webui with the manager

[tawny saddle]

maybe comfyUI is a playful abbreviation to Complicated Ui

[buoyant lava]

why exactly does putting (hands:1.2) in the negative prompt make the hands look fine

[mystic dirge]

Does anyone know why this doesnt launch? Its been "processing for 1800 seconds" and Im even overclocking my GPU

nevermind, I just uninstalled and reinstalled it

Dude. No. Stop shilling your cr ypto server crap.

[prisma moss]

because hands are overtuned (as in theres so much of the combinations of them) so this reduces the load

I had good experience with hands:1.2 in negs too, but only on some models

[onyx solar]

Beware: scammer. ⬆️

[opal crag]

hey guys, do u know any models when i can upload this image and make it like 3d model reference
2d to 3d type of model from civit ai

[brave frigate]

anyone know how to fix this?:

RuntimeError: mixed dtype (CPU): expect parameter to have scalar type of Float

?

idk man, seems sus

never had to do this

I ain't joining some random discord server, i trust this one and you're not a mod or anything as far as i can tell

also, i think @indigo vessel is a scammer so someone pls check that or ban him.

posted this sus message.

[fair oxide]

can someone please help im trying to compile Automatic1111 but i get this error:

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for tokenizers
Failed to build Pillow tokenizers
ERROR: Failed to build installable wheels for some pyproject.toml based projects (Pillow, tokenizers)

no idea whats going on im new to all this, so frustrating

[ornate elk]

What's your GPU?

[lime pivot]

Does anyone know whats the best tool for animating an anime style SD image ? Like if I have a character and I want to make him move. From what Ive heard, ppl say Luma is good

[prisma moss]

uh, kling, wai, huanyan

[west mural]

Any suggestions to queue several task with forge webui ?

[lime pivot]

thank u!

[brave frigate]

AMD Radeon RX 7900 XT

CPU is AMD Ryzen 7 7800X3D 8-Core

[ornate elk]

Ah okay then I guess you updated to the latest AMD Adrenalin Driver.
That needs you to manually update the zluda files to 3.9.5 in the .zluda folder.
Or you downgrade to Adrenalin 25.4.1

[brave frigate]

how do i update the zluda file?

[ornate elk]

Download the zluda 3.9.5 rocm6 zip (not nightly) files from here:
https://github.com/lshqqytiger/ZLUDA/releases/

Unzip the folder.
There make copy of the cublas.dll and the cusparse.dll and nvrtc.dll repaste them inside the folder.
Rename the copies to cublas64_11.dll and cusparse64_11.dll and nvrtc64_112_0.dll
Now copy all files from that folder into the .zluda folder inside the stable-diffusion-webui folder.

Then relaunch the webui-user.bat

[brave frigate]

I will do that rq

Also this:

[ornate elk]

Thx we will ban them

[brave frigate]

seems to work
compilation is in progress. Please wait...

[vocal burrow]

For what its worth there should be an option in discord to turn off DM during onboarding / whenever you enter a new discord. (they should be off by default)

[brave frigate]

Thx it works again!

You're an Angel 🩷

[sharp junco]

Hi! Why do my image look so Pale? When its generating it is colourful like in the 2nd picture, but in finale it's always Pale, colourless like in the 1st picture. Even when i clip skip=2

[vocal burrow]

btw you re using an SDXL model.
SDXL models are trained for 1024x1024, so your output should around 1024x1024 too (with more or less 25%).

[vocal burrow]

most likely you re not using the correct VAE

[brave frigate]

I use a x2 Hiresfix that brings em up to 1080 and 1920

so im within 540x960

[vocal burrow]

the upscaled resolution does nt really matter. You were using 512x512 in your previous screenshot. 540x960 is quite short in the x axis you might have wonky outputs.

[brave frigate]

Nah i have really good outputs been using it for quite some time now, want to see some of my outputs?

[ornate elk]

Perfect, no problem 🙂

[vocal burrow]

Nah it s okay. I m just warning you about your workflow. 540 is quite low for sdxl, it might work but it might also just output garbage, keep that in mind if it starts acting up.

[sharp junco]

Nah i used the same Vae as intended

[ornate elk]

can you show the full generation data? which webui do you use?

[azure magnet]

@ornate elk Hi, could you help me to upgrade comfyui portable?

[ornate elk]

Hey you can upgrade it with the Comfyui Manager

[azure magnet]

Don't work, it say comfy is already update to the last stable version but i'm at 3.27 where there is 3.41

[ornate elk]

And how did you installed Comfyui?

[azure magnet]

It's the portable version so i just download it

[ornate elk]

then go in the update folder and run the update_comfyui_stable.bat

[valid kettle]

Can someone give me some advice? I have a lot of 2d images i'm using as sprites in a game i'm making. But i want to run them all through an img2img model to apply a similar art style / filter over them. I'm trying to get a simple cartoony style applied to them.. Can anyone give me some advice on how i should do it?

[light tree]

I will help your work.

[valid kettle]

how?

[light tree]

You may use Stable Diffusion with ControlNet or IPAdapter(img2img with structure control)

Have you tried this?

[frail cobalt]

Hi, can anyone recommend the easiest/simplest way for scheduling/timing prompts and LORA's in comfy? thanks

[random turtle]

Did anyone run into this issue where if you have solid background (mainly gray / white / blue) the part of the image the runs through the k-sampler has a weird color shift. It only happens with solid background. I assume it's somewhere in a rgb/srgb color shift?

[prisma moss]

did you inpaint that thigh?

[dense viper]

I'm getting ModuleNotFoundError: No module named 'gguf' error after updating my webui forge in stability matrix, looked up on solutions related to this https://github.com/lllyasviel/stable-diffusion-webui-forge/issues/1134#issuecomment-2290925628 didn't work. it was working perfectly fine until this update

[dense viper]

seems like it's not only me getting that issue https://github.com/lllyasviel/stable-diffusion-webui-forge/issues/2943

[ornate elk]

Can you locate the forge folder?

Do you have a venv folder inside there?

[dense viper]

there's a venv folder inside, tried deleting it also didn't work

[ornate elk]

Don't delete it.
If you are in that folder where the venv folder is located do this:
In the Folder click into the File Explorer bar (not searchbar).
Type cmd and hit enter.
Then run the following two commands one by one:
venv\Scripts\activate.bat
pip install gguf

[potent sun]

is anyone here? I wanna discuss some training settings

Im getting ggod results but they could be better

[dense viper]

that worked, but now I'm running into another error AssertionError: Torch not compiled with CUDA enabled

[ornate elk]

Make sure your GPU driver is updated

A full cmd log would be nice too

[ornate elk]

Remove the launch arg --skip-torch-cuda-test and then delete the venv folder again.
Then relaunch.

You then also need to install gguf again like before

[dense viper]

Just fresh installed SD next and it's working fine, I'll see if I can reinstall forge webui to see if that works

[ornate elk]

What's your GPU btw?

[dense viper]

RTX4070ti, recently swapped from RX7600

I was lazy switching from amdgpu forge but it was working fine

[ornate elk]

Ah okay.
Would recommend reinstalling forge without the use of stability matrix.
It just makes everything more complicated

[light tree]

It has to be installed by "torch+cu+" version.

[dense viper]

I just reinstalled webui forge and all of sudden it's working now, thanks for the help anyway

[sharp junco]

What is the most beautiful model for Forge now?

for the work in colourful anime style, or fantasy style

for making dnd scenes with characters

i have 5080RTX

[mossy totem]

I keep pressing generate but nothing is happening

[ornate elk]

What does the cmd shows?

[mossy totem]

I have no idea

[ornate elk]

You can copy and paste it here

If there is no cmd window open the webui isnt running

[mossy totem]

aight

File "C:\Users\Owner\Desktop\stable-diffusion-webui\venv\lib\site-packages\torch\cuda_init_.py", line 769, in current_device
lazy_init()
File "C:\Users\Owner\Desktop\stable-diffusion-webui\venv\lib\site-packages\torch\cuda_init.py", line 298, in _lazy_init
torch._C._cuda_init()
RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx

AH

so am I missing whatever this is then?

[ornate elk]

Yea your missing the nvidia GPU driver.
Do you have an Nvidia GPU in your PC?

You can check that in the taskmanager under the performance tab, there under GPU 0 or GPU 1

[mossy totem]

I'm using a generic laptop, there is no secondary prosseser of anykind

[ornate elk]

Okay so provably only an Intel UHD onboard chip then

[mossy totem]

yeah

[ornate elk]

That means you can only use the very slow CPU mode with the webui.

Its not really fun to use that

[mossy totem]

that's fine, how do I set it to run in CPU mode?

[ornate elk]

You have to edit the webui-user.bat file.
Right click and edit.
At the line
set COMMANDLINE_ARGS=
you have to add:
--skip-torch-cuda-test --use-cpu all --no-half --precision full --opt-sub-quad-attention

Then save and close the file.
Then delete the venv folder.
Then relaunch the webui-user.bat

[mossy totem]

where do I find the venv folder?

[ornate elk]

Its in the stable-diffusion-webui folder

If its not there then you can ignore it and launch the webui-user.bat

[mossy totem]

if I tell you the name of the folder I think I should delete can you verify for me?

I found a folder that is litterlly just called Venv, it is a folder, not a file, this is what I'm supposed to delete?

[ornate elk]

Yes that folder

[mossy totem]

ok, so now I launched after doing that and it says this:
Creating venv in directory C:\Users\Owner\Desktop\stable-diffusion-webui\venv using python "C:\Users\Owner\AppData\Local\Programs\Python\Python310\python.exe"
Requirement already satisfied: pip in c:\users\owner\desktop\stable-diffusion-webui\venv\lib\site-packages (23.0.1)
Collecting pip
Using cached pip-25.1.1-py3-none-any.whl (1.8 MB)
Installing collected packages: pip
Attempting uninstall: pip
Found existing installation: pip 23.0.1
Uninstalling pip-23.0.1:
Successfully uninstalled pip-23.0.1
Successfully installed pip-25.1.1
venv "C:\Users\Owner\Desktop\stable-diffusion-webui\venv\Scripts\Python.exe"
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: v1.10.1
Commit hash: 82a973c04367123ae98bd9abdf80d9eda9b910e2
Installing torch and torchvision
Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/cu121
Collecting torch==2.1.2
Using cached https://download.pytorch.org/whl/cu121/torch-2.1.2%2Bcu121-cp310-cp310-win_amd64.whl (2473.9 MB)

is that ok? should I stop it?

[ornate elk]

Looks good, let it run

[mossy totem]

okie dokie

[ornate elk]

I'm off now as its late.
The webui should open up in the Browser when its done installing.
If not let me know.

[mossy totem]

it didn;t open

launch.py: error: unrecognized arguments: --precision-full
Press any key to continue . . .

and then it closed

[mossy totem]

.

[ornate elk]

Oh sry It has to be --precision full