🤝｜tech-support

[ornate elk]

Yes

But automatic does the samw

How does your image look like?

[austere spruce]

no, same result

wdym

[ornate elk]

If your not using a VAE the image might be dull or doesn't have much contrast

3060 is a good entry card for SD stuff

[austere spruce]

i'd say its somewhat dull i guess

but i dont think it differs if i use different model or image

[ornate elk]

Some models have inpaint versions of their model that could help

But first you should get a VAE file.
For realistic models its the 84000-mse VAE
And for anime the kl-f8-anime2

[austere spruce]

it like shows like this during generation and then literally flips to this

[austere spruce]

ok i'll try

[ornate elk]

Yea I've seen that behaviour before. But I'm not at the PC to do some quick test

[austere spruce]

hmm where do you put vae model

[vocal burrow]

#💬｜general-chat message

[nova ermine]

I'm trying to run an extension called SadTalker, but I've got this error

I've been told I'm missing a Python module called numpy, but I have no idea what to do.

[austere spruce]

looks like vae helped

[crystal nacelle]

yes, thank you for your reply. it was already disabled. maybe i need to update python? i tried updating but the venv is still referencing my old python

[fast hamlet]

hey guys, I am using ponydiffusion model and while its renndering out. no problem. onces its done. I see the final image then its starts artifacting like this

I dont know how to fix it

i redownloaded the model and disabled any LORA i was using

no clipskips or overrides

[nova ermine]

Try reducing cfg scale

[fast hamlet]

its only at 7

[nova ermine]

Try 4, the result may be unconsistent, but it may solve he issue

I had a similar problem a time ago, in my case I had too much weight in some promps and I had to reduce it. I don't know too much of this topic however.

[fast hamlet]

so maybe too many promps?

[ornate elk]

What's your python version?

[fast hamlet]

if thats the case, maybe I could reduce the amounnt and then refeed the ai my generated image and inntroduce more promps?

[ornate elk]

You need to use the sdxl vae

And dont use an 1.5 VAE for pony

[fast hamlet]

how do I know if I am using 1.5 VAE?

I Have the sdxl_vae.safetensors inn my VAE folder

[vocal burrow]

cf #🤝｜tech-support message

[fast hamlet]

but do I need to manually switch it in the ui?

sorry to be a bother, is it inn my settings?

[vocal burrow]

probably depends if you added the sd_vae to the quicksettings list

[umbral storm]

Hey guys. This issue has been plaguing me for quite a while. For instance the OpenPoseEditor tab next to the Extensions tab just vanished all of a sudden a couple of weeks ago.

None of the extension tabs show up anymore, not even new ones, like the Depth Lib menu.

I think I tried reinstalling openpose editor before, even a different version of it, but it didn't help at all.

I'm not sure if reinstalling all extensions on this client would help since I already tried installing extensions on a fresh client and it still did not show the OpenPose Editor for instance.

Anyone who's experienced this issue before?

[vocal burrow]

but again cf #🤝｜tech-support message everything VAE related should be explained there

[vocal burrow]

In your 3rd screenshot click "check for updates" let it do its work and then click "Apply and Restart UI"

[umbral storm]

Will do and update, thanks 😄

[ornate elk]

Hey, openpose editor is outdated and doesn't work anymore.
But it got integrated into controlnet.
You have to input an image.
Click on openpose and then on the Explosion icon.
Then you get and Pose and on the right an Edit button

[vocal burrow]

then reboot your webui and if the problem still happen please provide the full console log

[fast hamlet]

Forgive me for beinng slow, but i dont know what cf is? I clicked the link you put in the chat but it doesnnt direct me anywhere

[umbral storm]

That's good to know, thanks! It still sucks that Depth Lib does not show up for me tho lol

[vocal burrow]

cf <> latin for compare. basically mean "check this out". try to click multiple times. Discord can be kinda jank with messages links sadly

[vocal burrow]

VAE quick guide :
1/ What is a VAE.
It's a part of the stable-diffusion pipeline that encode/decode information from tokens to latent space and from latent spaces to pixels. Aka it transforms math statistics in pictures
2/ Where do I put my VAE ?

• VAE with .vae.pt, .vae.ckpt, .vae.safetensors extensions go into the models\Stable-diffusion folder
• VAE with .pt, .ckpt, .safetensors go into models\VAE
  3/ How do I use my VAE ? Three possibilities :
• Either you name it similar to another one of your model (eg : Anything-V3.0.safetensors + Anything-V3.0.vae.pt), by doing that it should automatically load the VAE when you load the associated model.
• You manually load your VAE by going to Settings -> Stable-Diffusion -> sd_vae and selecting your VAE
• You add an easily accessible VAE dropdown at the top of your page to quickly switch back VAE by adding sd_vae to your Settings -> User Interface -> Quicksettings list

There, I reposted it

[ornate elk]

Make sure your using the fork of depth lib and deleting the old one:
https://github.com/wywywywy/sd-webui-depth-lib

[fast hamlet]

actual legend man

thannk you

[umbral storm]

Thanks a lot! Will do this right now. Hope it will do the trick 😄

[ornate elk]

It will. Because the old version is not updated since 1 year

The new one should work.

[nova ermine]

I would thank you if you could help me with this

[umbral storm]

Thank you! You are an absolute legend 💙 🤜🏻 🤛🏻

[crystal nacelle]

Python 3.10.0

[ornate elk]

You need 3.10.11

[chilly knot]

can i run in google colab xl models?

[upbeat atlas]

It's RX 6900 XT. Btw what about the grainy images? I mean it's not so grainy that ruins the image, but it's still very noticeable, and the faces sometimes are distorted so I'm forced to use Adetailer. Any ideas on how to fix it?

[left jackal]

i keep getting this utf-8 cant decode byte 0x96 error. i reinstall python, reinstall my pc, changed training image, training name. still cant fix this. please help me

[upbeat atlas]

Quick question, can you? Isn't stable diffusion banned on google colab?

[ornate elk]

Do you have an example?
Sdxl is known to be grainier than 1.5

[chilly knot]

no i think

i found many tutorial

[upbeat atlas]

Ok so I just added a bunch of score tags and it somehow got a lot better, thanks as always

[hot spindle]

Gonna asnwer myself as I found the answer and some other people could be interested: yes, controlnet photomaker model is a .bin file that goes in models directory but the reason I was struggling to make it work is that (at least for Forge and maybe A1111) one needs to include the word "photomaker" in prompt

Bur on the same note: only the second IP-adapter option seems to be working for me, does the first one (face) requires a diferent model?

[hot spindle]

Again, answering myself 😁 : https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/178#discussioncomment-8572388

[waxen cradle]

hi, someonca can help me to change folders destination?

need change the path

what´s wrong here?

[sand grotto]

hi what would be the best .bat for my build

I7 9700k
3080
32GB ram

[ornate elk]

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

would be the best for 10gb vram

[ornate elk]

hey, it should look like this: (you just have to edit the path at the top to match yours)

also never install a webui inside the OneDrive folder

[sand grotto]

@ornate elk im getting this error creating an iamge Warning: field infotext in API payload not found in <modules.processing.StableDiffusionProcessingTxt2Img object at

[ornate elk]

can you show the full cmd log?

[sand grotto]

NotImplementedError: No operator found for memory_efficient_attention_forward with inputs:
query : shape=(1, 4104, 10, 64) (torch.float16)
key : shape=(1, 4104, 10, 64) (torch.float16)
value : shape=(1, 4104, 10, 64) (torch.float16)
attn_bias : <class 'NoneType'>
p : 0.0
cutlassF is not supported because:
xFormers wasn't build with CUDA support
flshattF is not supported because:
xFormers wasn't build with CUDA support
tritonflashattF is not supported because:
xFormers wasn't build with CUDA support
triton is not available
requires A100 GPU
smallkF is not supported because:
xFormers wasn't build with CUDA support
dtype=torch.float16 (supported: {torch.float32})
max(query.shape[-1] != value.shape[-1]) > 32
unsupported embed per head: 64

---

[tardy pasture]

can someone from stability confirm that it is not possible to disable the safety checker via the API? i couldn't see na option in the docs. our users are often generating content for kids so the safety check tends to trip when describing characters. i know we can use SD on other platforms with it turned off, but we're already integrated.

[ornate elk]

thats not the full cmd log.
what webui are you using?

auto1111?

[sand grotto]

yes

[ornate elk]

then please restart it and show me the full cmd text

[sand grotto]

found the issue folder

how can i map folder?

[ornate elk]

a models folder?

[sand grotto]

yes it seems its not detecting the lora when sear manually i dont see it

tried to recreate this

cant attach soo look up https://civitai.com/images/645644

[ornate elk]

loras are either for 1.5 or sdxl models. so make sure to have the right model version selctet as checkpoint

[ornate elk]

the lora used here is for 1.5 models

[sand grotto]

and i choose the realistic vision 6.0 b1

[ornate elk]

is the lora in models/lora ?

[harsh cove]

Hi everyone! I'm Katia, a student studying Information Systems Engineering. I'm currently working on a project and I'm interested in incorporating an API to generate images from text. I've heard about the Stability API, but I'm not sure how to get started, especially regarding obtaining an API key. Could anyone here guide me on how to integrate this API into my project? Any help or advice would be greatly appreciated!

[ornate elk]

hey for infos about the Stability.ai API checkout this:
https://platform.stability.ai/
https://platform.stability.ai/docs/getting-started

the second link covers the steps needed

[harsh cove]

Thanks! 🤩

[sand grotto]

is this normal :

WARNING:xformers:WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.0.0+cu118 with CUDA 1108 (you have 2.1.2+cu121)
Python 3.10.10 (you have 3.10.6)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details

[ornate elk]

you need to update your python to 3.10.11 64bit

download the installer and overwrite the existing python installation

then delete the venv folder and relaunch

[sand grotto]

how u mean byu cmd

?

[ornate elk]

no, with the installer from here (at the bottom)
https://www.python.org/downloads/release/python-31011/

[sand grotto]

delete the venv folder inside a1111

[ornate elk]

yep

it will get recreated after launching the webui-user.bat again

and then its using the new python and xformers

[sand grotto]

6gb wow this folder is big by default?

[ornate elk]

yes, there are the core files stored, torch etc

[nova ermine]

how to update numpy?
I'm having an issue and apparently that would solve

[ornate elk]

i tried sadtalker on my auto1111 install, getting the same error

[nova ermine]

Oh

[ornate elk]

the devs havent updated the extension for the webui 1.8.0

so no fix

[nova ermine]

There's no way of using it?

[ornate elk]

you can checkout the issues tab of their github site:
https://github.com/OpenTalker/SadTalker/issues
maybe someone has a fix, but i dont think so

[nova ermine]

I saw a tutorial of a man who used it, but he uses as an individual tool (not on the webui)

However, I got lost :'(

[ornate elk]

you cna try their online demo:
https://huggingface.co/spaces/vinthony/SadTalker

[nova ermine]

I've been having errors with that

So, do you consider I would have problems running it locally outside automatic1111?

[ornate elk]

i dont know

you can try to set it up as standanlone tool

[sand grotto]

how does one mapp folders in the bat file?

[ornate elk]

for models and loras?
--ckpt-dir "Path/to/models/Folder"
--lora-dir "Path/to/loras"

[nova ermine]

Actually I got that running it outside Automatic1111. I don't know if it's the same error

[ornate elk]

could be a python error

[sand grotto]

and others?

vae , embedding a

ect

[ornate elk]

--embeddings-dir
--Esrgan-dir
--vae-dir
--controlnet-dir (extension neede first)

[lost crescent]

I have 55 LoRAs in ComfyUI. I am using the same LoRA folder in A1111 - but it will only display 46 of the same LoRAs?!?!?!? Why?

[ornate elk]

maybe some are for sdxl

[lost crescent]

ALL are for SDXL - but 10 or so LoRAs fail to load in A1111 - which is very frustrating 😄

I understand that if you're using SD2 - only SD2 LoRAs will load; and if using SDXL, then the SDXL LoRAs should load ...

... but

[nova ermine]

I finally got this. It fixed the error and I could be able to run sadTalker. I just send this in case you are interested in knowing

[ornate elk]

oh good to know, is that for the extension or the standalone?

[nova ermine]

I don't know what is standolone

That fixed SadTalker. However, I'm running it outside the webui

I don't know if it fixes the same inside the webui. I've got to much headache after dealing with this for all the day long so I'm not considering to check it

[sand grotto]

for some odd reason it wont go deeper in o the folders im dividing loras in human fantasy cars buildings and abstract

[sonic frigate]

heeeyaaa, where can I find a local to download

[ornate elk]

You can just create subfolders

They work

[ornate elk]

Check the pinned messages in this channel for the right guide

[sonic frigate]

oke

ty

[strong kestrel]

help it sayis connection timed out right after starting in on the webui

[sonic frigate]

hey, one question, it necesarily need to be on c: or can be on an external memory

?

[willow flare]

What?

[ornate elk]

It can be on any drive. But git and Python preferable on C

[sonic frigate]

aa ye ye

ty

[ornate elk]

SDXL models are trained on a base resolution of 1024x1024

So try that

[ornate elk]

Make sure you didn't closed the cmd window

[strong kestrel]

i didnt

its still open

it errors out right after loading the page

[ornate elk]

Any error in the cmd?

[strong kestrel]

where do i see?

[ornate elk]

In the black window

[strong kestrel]

idk you tell me, no clue what to look for

[ornate elk]

What's your GPU?

[strong kestrel]

rtx 3080 laptop gpu

[ornate elk]

@strong kestrel
okay, then do the following:
right click and edit the webui-user.bat
at the line COMMANDLINE_ARGS=
you need to add: --xformers --medvram-sdxl --no-half-vae
then save and close.
then delete the venv folder and relaunch the webui-user.bat

[vocal burrow]

it's stating everything you need to know

(you can also delete venv instead like cs1o suggest)

[strong kestrel]

what?

[strong kestrel]

okay i'll try that

[strong kestrel]

how do i edit? when i open it, it just starts it up

[strong kestrel]

ah oke, like this?

[ornate elk]

xformers is missing a -

--xformers

[strong kestrel]

mhm i added

it still errors out

[ornate elk]

how much RAM do you have?

[strong kestrel]

32gb

this was never a problem, it worked fine 2 months ago, but now it doesnt

[sonic frigate]

here can I add a img to use as a base?

to edit a image?

[ornate elk]

any error now in the cmd?

[strong kestrel]

lots of warnings now

[ornate elk]

have you deletet the venv folder?

[strong kestrel]

no

[ornate elk]

then do that please

[strong kestrel]

now its saying creating vnev so its adding it again though

[ornate elk]

yes thats good

[strong kestrel]

oh okay

[ornate elk]

let it load

[strong kestrel]

will do

[ornate elk]

nope

[sonic frigate]

and the local one?

[strong kestrel]

it still times out TwT

now the whole browser page says website cant be reached

help what do i do

[ornate elk]

what browser are you using?

make sure the webui is whitelisted in any browser adblocker

[strong kestrel]

operagx, but tried it in chrome too and there it says ERR_CONNECTION_REFUSED

[strong kestrel]

yea i already tried that

when i try to reload this happens

[ornate elk]

and the cmd is still open?

[strong kestrel]

yes

always says errored out

help

why doesnt it work?

this is so frustrating

would a reinstall fix this?

hello??

pls help

[ornate elk]

You could try a reinstall with my guide in the Pinned Messages of this channel

[strong kestrel]

but would that fix it

[ornate elk]

the error is strange, do you have any third party antivirus software installed?

[strong kestrel]

nope just the windows defender thing, even turned off the firewall but no change

[ornate elk]

have you also restarted the Laptop?

[strong kestrel]

oh yes, more than once

oka yi just installed it on my macbook and the cmd window looks a lot different

[ornate elk]

check if your C: drive has enough space left

[strong kestrel]

it does

[ornate elk]

okay then try deleting the config.json and ui-config.json

[strong kestrel]

ookay no errored out so far...

[karmic crown]

on mac? 🤣

[ornate elk]

No

[strong kestrel]

no

ah nvm now it errored

[karmic crown]

i see macos here, i didn't read the whole conversation

[strong kestrel]

on the last line it says "press any key. . ." and when i do it closes instantly, that doesnt happen on my mac

[karmic crown]

you are trying to get the same thing on windows that you have on mac, and mac works?

[strong kestrel]

No if you dont know whats going on then leave

[karmic crown]

i am trying to figure out if you need help with mac or not

[strong kestrel]

no

[karmic crown]

if you need help with windows CS1o can help you more than me
but if you need help with mac i can help more than him

[strong kestrel]

on my windows laptop which i used for stable diffusion before, it worked perfectly then, alyways says this all of a sudden and idk why

was just trying to see if the same thing happens on mac too

[karmic crown]

ok, i just read all

so you are getting timeout on windows, with local connection

[strong kestrel]

yes

seems like it

[karmic crown]

does ui open now or not? like on this picture, if it does, check what you have on the bottom

this part

[strong kestrel]

it always opened

[karmic crown]

please post this if you can see it

[karmic crown]

i do not see the versions there

[strong kestrel]

?

[karmic crown]

this part

on the bottom

[fast hamlet]

getting a lot of artifacts, using pondydiffusion with lora, and the SDXL_VAE vae

eyes are not defined, air is bugging out. colors seemed warped

[karmic crown]

hmm, gradio version looks fine

[fast hamlet]

im a noob to using stable diffusion. Any help would be appreciated

[karmic crown]

@strong kestrel one stupid question, have you tried with some other browser like edge?

some people had problems with adblokers

i am not sure if opera has adbloker or not

that is the only browser that i do not have and i have 5 of them

[strong kestrel]

it does but i have disabled it for this

[ornate elk]

You can try update your python to 3.10.11 64bit
Then delete the venv and relaunch the webui-user.bat
If its still errors out. I would try a clean reinstall

[karmic crown]

i do not think that would help

[karmic crown]

can you please try edge too, so we can confirm that the problem is not adblocker

[karmic crown]

thanks

since that is windows, grap probably does not work, but netstat should work (i hope)

❯ netstat -an

[strong kestrel]

?

[karmic crown]

open another command prompt

and try to type that

[strong kestrel]

what?

[karmic crown]

netstat -an

[strong kestrel]

???

command prompt?

[karmic crown]

yes CMD window

[strong kestrel]

its already open

[karmic crown]

another one

[strong kestrel]

and when i press anything it closes

how?

[proven finch]

try https://localhost:7680

[strong kestrel]

doesnt work

[karmic crown]

Press Windows+R to open "Run" box. Type "cmd" and then click "OK"

[proven finch]

huh

does the share link work?

[strong kestrel]

okay i have run it but what now?

[strong kestrel]

?

[karmic crown]

type netstat -an and press enter

[strong kestrel]

i just told you i did

[proven finch]

viking what are you tring to do

[karmic crown]

❯ netstat -an | grep 7860
tcp4 0 0 127.0.0.1.7860 . LISTEN

i am on mac

i want to see if port is opened or not

since he has no grep i am trying to see if he can get the full port list and to try to find it by search in notepad

[strong kestrel]

i'm not a he

[proven finch]

open webui-user.bat ; add --share to the cli args; save and re-run

[karmic crown]

but i am not sure if netstat works on windows or not

[proven finch]

or .sh

[proven finch]

it does

[strong kestrel]

i have it opened btu cant add anything to it cuz it instantly closes when i press a key

[proven finch]

oh open it with notepad

to edit it

[strong kestrel]

whats cli agrs

[fast hamlet]

can anyone help please?

[proven finch]

model issue

[karmic crown]

he has a problem with access from local machine to local port

[strong kestrel]

IM NOT A HE

[proven finch]

share should fix it

[karmic crown]

sorry SHE

[fast hamlet]

but this is a widley used modle. i havent seen many complaints

[proven finch]

could also be prompt issue

OH

remove git pull

(also is medvram-sdxl a real thing?)

[proven finch]

yea

[strong kestrel]

idk someone told me to add it

[proven finch]

okay try running it now

[strong kestrel]

do i need to delete that ve- thing folder again?

[karmic crown]

yes it is great parameter

no

[proven finch]

venv? no

[strong kestrel]

okay

[proven finch]

never delete venv*

[strong kestrel]

i was kept told to delete it

[proven finch]

does it say "Press key to stop blah blah.."?

[strong kestrel]

yep

timed out again

[proven finch]

huh.

something's wrong with the internet connection

also

[strong kestrel]

no its not

[proven finch]

do you have any other terminal open

[strong kestrel]

otherwise i wouldnt be talking

and it worked fine a few months ago

[strong kestrel]

no

[karmic crown]

venv can be safely removed if that is needed, because it will be created on next run

she didn't do anything wrong

[proven finch]

yeah uh

at this point id say to uninstall automatic and install comfy but

something must be wrong with ports

i think

[strong kestrel]

then i have to install all the lil things for my model again T^T and that was such a pain so i dont want to do that, and i would lose all my images in the output folders

[karmic crown]

@strong kestrel do you have some 3rd party antivirus, like mcafee?

[strong kestrel]

no

wait it only times out when changinbg to a specific model

[proven finch]

oh no comfy uses the model you already have, but its for masochists

[strong kestrel]

thats new

[karmic crown]

adbloker - checked
wierd 3rd party antivirus - checked
...
i have no idea what else could make this problem 😦

[strong kestrel]

the model i want to use times it out though wtf

[proven finch]

Windows.

can we swear in this server

[onyx solar]

fuck no

[proven finch]

hey so basically windows says that you're fucked

[strong kestrel]

excuse me?

[proven finch]

tldr: windows lets the webui launch

but when changing models

it uses too much ram

and gets killed

[strong kestrel]

so how do i fix it?

[proven finch]

define the base model as the one you want to use

either with some cli arg

[karmic crown]

i have an idea

[strong kestrel]

????

[ornate elk]

ohh okay you tried a yiffymix, the furry models are often broken

[proven finch]

or by delting all other models

could also be too little vram

[proven finch]

dw ill guide you

[strong kestrel]

could do that, i dont need them anyways, just same model just previous versions

[proven finch]

whats the model name (like model.safetensors, cirno_oven.safetensors)

[strong kestrel]

hey it works now wtf

[ornate elk]

delete the yiffy model and restart

[strong kestrel]

i deleted the precious versions

[strong kestrel]

no nice try though

[ornate elk]

but please remove --share from your webui-user.bat

its not needed

[karmic crown]

@strong kestrel try this, go to https://github.com/lllyasviel/stable-diffusion-webui-forge and download >>> Click Here to Download One-Click Package<<<
unpack it in the new folder, and see if that works
if that works, your windows is fine, if not, you have a serious problem with Windows

[proven finch]

yea that too

if it works it works

[strong kestrel]

as i just told you, it works now, i deleted the previous versions of the same model

[proven finch]

this could speed it up

its mainly for mem efficiency

which you dont seem to need

[karmic crown]

that is not important now
if forge launch windows works fine
if not windows is a problem

reinstalling a1111 from scratch (including python) will not fix anything if windows is a problem
i am trying to save her time
this package just need to be unpacked and launched

and we will know what is the problem

[strong kestrel]

well thank all you lol what a frustrating issue

[karmic crown]

please try this, so we can tell you what to do next if that works / not work

[strong kestrel]

but didnt yoinked say i didnt need it?

[ornate elk]

good that it works now, dont forget to remove the --share
but leave the other args in the webui-user.bat, they are for the best performance

[strong kestrel]

okay!

[karmic crown]

wait, your webui now works? and you can access it without timeout?

[strong kestrel]

i said that like 3 times now

[karmic crown]

in that case forget forge

[strong kestrel]

haha ye oke

[karmic crown]

i thought you still have the same error

[strong kestrel]

luckily not, guess just too many models installed

[sand grotto]

question:

what does it mean : half-length portrait:1.2 << the numbering sorry english is not main tongue

[proven finch]

number mean how important the text before is

e.g.
(egg:0.4) not a lot of egg
(egg:1.7) das a lot of egg

[noble belfry]

-egg. thats a neggative 😛

[sand grotto]

thnx , what about painter how do you choose what alter in the generated image?

is that a thing? does it normally take to laong to generate i mena it has been 40 min

[noble belfry]

what GPU do you have for your computer?

[sand grotto]

3080 evga 10gb

shared memory ... 16gb

[noble belfry]

what resolution images are you making? 1024x1024? how many steps? 30? are you upscaling it?

[sand grotto]

915x1152

sampling steps 30 sorry

[noble belfry]

it really should not be taking 40 minutes.

40 seconds possibly.

maybe you're running without cuda?

[waxen ore]

just got this after updating compfyui just now? any info?

[ornate elk]

What 40mins?

[sand grotto]

i think its because of the 30 steps

[sand grotto]

this is my bat:
@echo off

set PYTHON=
set GIT=
set VENV_DIR=
set COMMANDLINE_ARGS= --autolaunch --update-check --xformers --medvram-sdxl --no-half-vae

git pull
call webui.bat

[sand grotto]

o it is using th gpu

there was an error before creating the image:

*** Error executing callback app_started_callback for F:\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\api.py
Traceback (most recent call last):
File "F:\stable-diffusion-webui\modules\script_callbacks.py", line 153, in app_started_callback
c.callback(demo, app)
File "F:\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\api.py", line 60, in controlnet_api
controlnet_input_images: List[str] = Body([], title='Controlnet Input Images'),
NameError: name 'List' is not defined

[waxen ore]

figured it out ok it seemed to be from an update to IF Prompt to Prompt Node from ImpactFrames Nodes..

[ornate elk]

what are your image settings?

[ornate elk]

30 steps should take seconds for your gpu

[noble belfry]

Skully — Today at 22:16
915x1152
sampling steps 30 sorry

has to be on gpu or something

[sand grotto]

it stopped

Warning: field infotext in API payload not found in <modules.processing.StableDiffusionProcessingTxt2Img object at 0x000001C1B6FC4910>.
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:13<00:00, 2.22it/s]
tiled upscale: 100%|███████████████████████████████████████████████████████████████████| 35/35 [05:59<00:00, 10.26s/it]
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [01:18<00:00, 2.60s/it]
Loading VAE weights specified in settings: F:\stable-diffusion-webui\models\VAE\anything-v4.0.vae.pt3:45<00:00, 2.51s/it]
Applying attention optimization: xformers... done.
VAE weights loaded.
Total progress: 100%|██████████████████████████████████████████████████████████████████| 60/60 [59:35<00:00, 59.59s/it]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 60/60 [59:35<00:00, 2.51s/it]

[old parrot]

Any way to make SD run offline with no internet connection?

[ornate elk]

It should work without internet

[old parrot]

Yeah I realized after it's because I was using gradio when my internet bugged out lol

Appreciate you

[fast hamlet]

In queue: 1/1

it seems i am waiting forever?

i do I just get to the next render I want?

[ornate elk]

Check the cmd Window for any errors

[ornate elk]

Dont use 0 hires steps, set them to 10 if you use hires fix

Also dont enable the refiner

[fair oxide]

My a1111 is only producing black images, its never been a problem before. How do I fix it

[ornate elk]

What model are you using?

[fair oxide]

Like the checkpoint?

GoofyAI v3

[ornate elk]

And what's your GPU and what's inside your webui-user.bat

[fair oxide]

6gb I think

And it's running a different model now

Just taking forever

[old parrot]

Use Stable Diffusion Forge instead, I have 6gb VRAM as well and it runs like shit on A1111

[ornate elk]

Then you need the performance arguments in the webui-user.bat

[fair oxide]

i have them

[ornate elk]

What's your gpu model?

[fair oxide]

i have no idea

[ornate elk]

Check in the taskmanager

Under performance tab

GPU

[fair oxide]

this??

[ornate elk]

Okay then remove --disable-nan-check --no-half --always-batch-cond-uncond

[fair oxide]

alright

[ornate elk]

These are not needed for your GPU and only cause issues or slowdowns

[kind idol]

trying to run stable cascade and i get a error "module diffusers has no attribute StableCascadeUNet"

anyone run into this issue

[fair oxide]

it wooorkkkeedd

thank you~

[ornate elk]

Perfect, np 🙂

[fair oxide]

back in working order wooo

[ornate elk]

Wow looks awesome!

[crystal nacelle]

Guys we need to all like start a donation box and get something for our boy @ornate elk coz this man has literally been carrying this discord server. Maybe a 3D printed luffy delivered to his place or something nice just to show our appreciation

[fair oxide]

hes a g, id donate, im only here cuz he gives good advice on how to fix my a1111

[crystal nacelle]

Same

[quasi ginkgo]

Hey yall, got an issue with the Civitai browser+ extension not working properly. When I hit search, it loads for a few seconds and all the input elements return an "error" tag. I'm getting an 101 error in the Inspect Element console/network tab and no errors in the SD console

Tried some googling and haven't found anything

If this is the wrong discord for this issue, I'd appreciate if I could be pointed to the appropriate one

[vocal burrow]

what does the console log looks like ?

[quasi ginkgo]

Inspect Element console or SD?

[vocal burrow]

sd

[quasi ginkgo]

hold

[vocal burrow]

dump the full log in here (properly formatted or as a .txt)

[quasi ginkgo]

Where might I find this log?

[vocal burrow]

the log is not saved anywhere, it's displayed in the console (black window thing with white text in it)

[quasi ginkgo]

Ah thought there might be a log output file somewhere, I'll make a .txt rq

I skimmed it and couldn't see anything relevant. Hitting search multiple times just gives the same 101 network errors and nothing in the SD console

[vocal burrow]

looks good so far, did you trigger the error before sending the log ?

[quasi ginkgo]

Yessir

[vocal burrow]

are your extensions fully updated ?

[quasi ginkgo]

Yup, updated them before I came here

[vocal burrow]

hmmmm can you screenshot your webui so I can see for myself what it looks like

[quasi ginkgo]

Sure thing

[vocal burrow]

the url is whitelisted in any kind of adblocker you're using ?

[quasi ginkgo]

I'm using UBlock Origin, not sure if its whitelisted

[vocal burrow]

might as well give it a shot

you might also try with a different browser (something clean of any extensions)

[quasi ginkgo]

Disabled the adblocker completely on my browser, no difference

[quasi ginkgo]

I'll give'r a shot

[vocal burrow]

time to dust off that edge browser :p

[quasi ginkgo]

no kidding

I guess I uninstalled chrome at some point but windows kept the shortcut in the start menu lol?

No dice on Edge

[vocal burrow]

damn

booting up mine to see if I can reproduce the issue

[quasi ginkgo]

10-4

Gonna try to install the extension directly from the github url, might do something

[vocal burrow]

I doubt it, everything looks clean :/

are you using some vpn ?

or should you be using some vpn (china, etc) ?

[quasi ginkgo]

I have vpns installed, but they're off

[vocal burrow]

and you can go on civitai using your browser just fine ?

[quasi ginkgo]

Yup

[vocal burrow]

when in doubt.... delete venv and try again ?

[quasi ginkgo]

damn bro are we really at that stage 😦

[vocal burrow]

We don't have much to work with :/ no logs whatsoever.
You can try to up the log level with command line arguments --loglevel DEBUG and we'll see if something interesting shows up

[quasi ginkgo]

Hold

I changed tabs and and something populated

[vocal burrow]

(still trying to boot up mine, I'm in the middle of something so it takes time)

[quasi ginkgo]

all good brotherman

[vocal burrow]

Theory : civitai has just refactored its categories for content rating system (yet again) and that's messing up the extension

if that's it then I should be able to reproduce the error easily

and you'll have to wait for the extension to be updated :p

[quasi ginkgo]

Saw that, figured it might have something to do with it but just wanted to make sure. I literally JUST installed the extension today so idk if I just had some bad settings or something

[vocal burrow]

try to turn off (or on) any kind of setting related to nsfw and see if it fixes it

[quasi ginkgo]

Already tried toggling that in the filter button, no changes

I got a clean install going from a different folder, if the issue persists on that instance then I'm just going to assume its an extension issue and go back to importing loras and checkpoints how I used to, like a fucking CAVEMAN

[vocal burrow]

https://github.com/BlafKing/sd-civitai-browser-plus/issues/249

yup sounds like it broke it

[quasi ginkgo]

29 min ago, I checked the discussions on there probably an hour ago

fml

[vocal burrow]

yeah I did when we started talking about it too :p

[quasi ginkgo]

maaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaan

[vocal burrow]

at least we came to the same conclusion

[dense sandal]

who doens't comfyUI create a new node instead of breaking the old one...

All my workflows are broken now

how can I revert to pre IPadapter node breaking update?

[quasi ginkgo]

@vocal burrow ty for taking time to help me tho man, I appreciate it

[steady compass]

in txt2img is it possible to make hires fix appear at the bottom and not at the top i want it to run last when generating images

[frail verge]

Has anyone had any issues with AUTO1111 lately where the web UI will literally not respond until it's reloaded?

• Clicking generate won't do anything
• PNG Info tab never loads
• Extensions tab won't load.

Tried different browsers (Chromium and Firefox Developer Edition - with extensions fully disabled & without) and the issue persists

No errors either - unless you count the web console errors that don't seem helpful at all

[split forge]

does anyone has an idea how to install a model such as this without a "model file"? ( https://huggingface.co/stablediffusionapi/anything-v5) i have a vague idea that change some texts and change the suffix, but i desperately need more instructions. xx

[solemn moon]

try delete venv folder and rerun

[frail verge]

ive already tried a fresh reinstall :(

[dreamy carbon]

From our old conversation. I would have to download "wildcardX-XL" and put it in model?

[solemn moon]

@brave cove
What’s your question?

[brave cove]

I want to generate my first image, how do I find the menu entry

[solemn moon]

I’m not sure what you mean by “menu entry” 🤔

[brave cove]

I mean I want to generate my first image, how should I do it

[solemn moon]

Check out some tutorials on YouTube
Either Automatic1111 or ComfyUI 😎🫶

[brave cove]

ths

[ornate elk]

Into models/stable-diffusion

[sudden hawk]

Hello, when i generate image it take me sooo much time (until 1h /1h30) while i do a ratio with 512 in width and 680 in height. Is it normal ? Or i should remove my parameters? I have nvidia rtx 3070

[solemn moon]

teh best thing is to paste the image parameters file, (npt sure if that wording), the text file with all the info, then people can check properly for you

[solemn moon]

paste the text file, cant see model etc

also whats in your command args?

[sudden hawk]

is this the text files ? And wym by command args ? (sorry english is not my maternal lgge and i'm not goog in informatic)

[thin jewel]

Hello.
I am using "Search and Replace" from Stable Image API.
Yesterday everything worked well, but today I'm receiving errors wih same prompts:
'{"error":"InferenceError: {\"code\":\"inference_error\",\"message\":\"Internal Server Error\"}"}'
What's wrong?

[ornate elk]

What's in your webui-user.bat?

[ornate elk]

Right click the webui-user.bat and edit it

At the line COMMANDLINE_ARGS
You need to add --medvram-sdxl to the other args

And remove --autolauch, it should be --autolaunch

[sudden hawk]

Thank you so much 🙏

[ornate elk]

It should be much faster now

@sudden hawk important is to say to always launch the webui-user.bat
And not the webui.bat

[sudden hawk]

Ok thanks again 🙏

[dreamy carbon]

tyvm

[lost crescent]

In ComfyUI, what is NDI?

[olive plinth]

Hi ! I alway get this Js issue at start of Stable Diffusion. I can't generate nothing until reload SD client by refresh web page. Anyone have get the same issues ? (I join my list of extensions too)

[ornate elk]

Hey, have you tried an other browser?

Also have you updated your extensions?

[olive plinth]

@ornate elkextention yes, browser no... I use Opera, maybe I must try with chrome

[ornate elk]

Yes, and also check the cmd for any errors when starting sd

[sharp burrow]

Hi, anyone using Stability.Ai with Make.com?

[gritty ridge]

Anyone training models locally on MAC ? Lora/Dreambooth- Kohya_ss is having issues and I know no other trainer…

I actually can not debug gui_sh or find out what is causing crash

[chilly knot]

when i try use canny

\torch\serialization.py", line 1112, in load_tensor
storage = zip_file.get_storage_from_record(name, numel, torch.UntypedStorage)._typed_storage()._untyped_storage
RuntimeError: PytorchStreamReader failed reading file data/4: invalid header or archive is corrupted

open pose works in control net

canny give me this

[vocal burrow]

try to redownload the canny model ?

with just that snippet of error, it seems like so

[chilly knot]

i do it now but where i get orig model

i download 5 gb model canny pth is correct?

[vocal burrow]

that seems like a lot

you can get difference controlnet model at 700mb https://civitai.com/models/9868/controlnet-pre-trained-difference-models

(those are to be used along sd1.5 models)

[chilly knot]

thx

i have present for you for help in general with images\

2 watermelon rats

super().init(torch._C.PyTorchFileReader(name_or_buffer))
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

with differrent canny model

[vocal burrow]

full log please

[chilly knot]

Error running process: :\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py
Traceback (most recent call last):
File ":\Stable diffusion\stable-diffusion-webui\modules\scripts.py", line 718, in process
script.process(p, *script_args)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 1143, in process
self.controlnet_hack(p)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 1128, in controlnet_hack
self.controlnet_main_entry(p)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 894, in controlnet_main_entry

[vocal burrow]

no, the full log please. everything written in the console.
copy paste it in a .txt and drop said .txt in here

[chilly knot]

ok

Error running process: :\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py
Traceback (most recent call last):
File ":\Stable diffusion\stable-diffusion-webui\modules\scripts.py", line 718, in process
script.process(p, *script_args)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 1143, in process
self.controlnet_hack(p)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 1128, in controlnet_hack
self.controlnet_main_entry(p)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 894, in controlnet_main_entry
model_net, control_model_type = Script.load_control_model(p, unet, unit.model)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 367, in load_control_model
control_model = Script.build_control_model(p, unet, model)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py", line 395, in build_control_model
state_dict = load_state_dict(model_path)
File ":\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\utils.py", line 21, in load_state_dict
state_dict = unsafe_torch_load(ckpt_path, map_location=torch.device(location))
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\torch\serialization.py", line 797, in load
with _open_zipfile_reader(opened_file) as opened_zipfile:
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\torch\serialization.py", line 283, in init
super().init(torch._C.PyTorchFileReader(name_or_buffer))
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

full

i think install torch not right

[vocal burrow]

that s not the full log

there s tons of stuff before that

include everything, scroll all the wy to the top

[chilly knot]

oh ok)

[vocal burrow]

if there s too much, restart your sdwebui trigger the error again and copy paste that log

[chilly knot]

Launching Web UI with arguments: --theme=dark --xformers --medvram --no-half-vae
: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable TF_ENABLE_ONEDNN_OPTS=0.
WARNING:tensorflow:From :\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.

Style database not found: :\Stable diffusion\stable-diffusion-webui\styles.csv
ControlNet preprocessor location: :\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\annotator\downloads
2024-03-27 13:25:39,457 - ControlNet - INFO - ControlNet v1.1.440
2024-03-27 13:25:39,804 - ControlNet - INFO - ControlNet v1.1.440

[balmy lily]

ahoy dank memers, im having some trouble installing stable diffusion, with this version
https://github.com/comfyanonymous/ComfyUI
while i was unzipping it, it gave some warnings that some specific files were missing, but when i used 7zip to unzip it it didnt give those missing file errors, so im not sure if thats the cause of the problem is or not. but i put some models in the checkpoints folder and a vae file in the vae folder, but the problem is when i launch it with run_nvidia_gpu.bat, and then the ui opens up in my browser, when i try to generate an image with the "queue prompt button", it just pops up a window that says "reconnecting..." in red letters.

and with this version of stable diffusion
https://github.com/AUTOMATIC1111/stable-diffusion-webui
also with models/checkpoints i put in, when i open the with the bat script, it opens the ui, but i keep getting red error windows that says "error. connection errored out"

my computer is pretty powerful btw so thats not the problem

[vocal burrow]

1/ there s supposed to be stuff before that
2/ you didn t trigger the error, so the actual error log is not there

here s what a full log looks like #🤝｜tech-support message

[chilly knot]

problem starting

2024-03-27 13:27:53,533 - ControlNet - INFO - unit_separate = False, style_align = False
2024-03-27 13:27:53,744 - ControlNet - INFO - Loading model: control_sd15_canny [fef5e48e]
*** Error running process: :\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\scripts\controlnet.py
Traceback (most recent call last):
File ":\Stable diffusion\stable-diffusion-webui\modules\scripts.py", line 718, in process
script.process(p, *script_args)

[vocal burrow]

ok no full log then....

make sure your extensions are updated

and make sure you re not using sd15 controlnet model along with a sdxl model (or vice versa)

[chilly knot]

hm

[ornate elk]

What extensions do you have installed?

Tensorflow error reminds me of TensorRT extension

[vocal burrow]

maybe something got corrupt during the download. Can you try to get them again ? Do you have logs for any of them ?

[balmy lily]

ive tried the downloads and unzippings and installations of them in difgferent ways and multiple times

[ornate elk]

When you launched the webui-user.bat and the ui opens, don't close the cmd

[balmy lily]

this is the log for just nows opening of automatic1111

Version: v1.8.0
Commit hash: bef51aed032c0aaa5cfd80445bc4cf0d85b408b5
Launching Web UI with arguments:
no module 'xformers'. Processing without...
no module 'xformers'. Processing without...
No module 'xformers'. Proceeding without it.
Calculating sha256 for E:\sd.webui\webui\models\Stable-diffusion\model.ckpt: Running on local URL:  http://127.0.0.1:7860

To create a public link, set `share=True` in `launch()`.
Startup time: 17.9s (prepare environment: 3.9s, import torch: 5.4s, import gradio: 1.5s, setup paths: 1.7s, initialize shared: 2.6s, other imports: 1.0s, load scripts: 1.1s, create ui: 0.4s, gradio launch: 0.3s).
cc6cb27103417325ff94f52b7a5d2dde45a7515b25c255d8e396c90014281516
Loading weights [cc6cb27103] from E:\sd.webui\webui\models\Stable-diffusion\model.ckpt
Press any key to continue . . .```

i mean xformers is prominent looking

[chilly knot]

roop control net open pose

[ornate elk]

Seems like you followed an old tutorial for that install

Please follow my install guide for automatic1111:
#🤝｜tech-support message

[vocal burrow]

no that s normal, that s because you re not using --xformers command line arguments. What s not normal is the "Press any key to continue . . ." which indicates that the program has terminated somehow without any errors.
What s in your webui-user.bat ? Are you using webui-user.bat to run sd ?

[balmy lily]

i did at first, but this installation i just did the instructions on the github

Download sd.webui.zip from v1.0.0-pre and extract its contents.
Run update.bat.
Run run.bat.```

[vocal burrow]

oh yeah, no

[ornate elk]

Yes thats not recommended

[vocal burrow]

don t use the zip if possible

check the pinned install guide

[balmy lily]

how should i uninstall my current 3.10.6 version of python most throughly

[ornate elk]

Not needed

You can upgrade it

[chilly knot]

i see in console ControlNet - INFO - ControlNet v1.1.440
ControlNet - INFO - ControlNet v1.1.440

double text\

its ok?

[ornate elk]

Thats okay I guess, when does the error appear? Also only when using canny model?

[chilly knot]

yes but i think i must see preview with open pose

but i dont see

[ornate elk]

Can you screenshot your controlnet settings?

[vocal burrow]

make sure to delete your venv folder (if there is any) after making any down/upgrade of your python

[vocal burrow]

and what model are you using it with ?

[chilly knot]

dreamshaper

[vocal burrow]

which one

[chilly knot]

8

for 1.5

[vocal burrow]

ok so it should be fine

[vocal burrow]

ok so let s try delete your venv folder and relaunch (include the full log once it s done)

maybe something installed incorrectly

[chilly knot]

if i use this setting it put pose from my image to generate with promt?

[vocal burrow]

it should yes

you can click the explosion logo to see a stickman preview of the pose

[chilly knot]

in right window must be stick man?

i think pose readed

[chilly knot]

in stable dif folder?

[vocal burrow]

Yes and sorry but I can't read those runes

[vocal burrow]

Yes

[chilly knot]

prewiev not work

i have same error with open pose when i chose preprocessor

if i use stickman without preprocessor it works

this problem

Using preprocessor: openpose_full

• ControlNet - INFO - preprocessor resolution = 512
  Downloading: "https://huggingface.co/lllyasviel/Annotators/resolve/main/hand_pose_model.pth" to :\Stable diffusion\stable-diffusion-webui\extensions\sd-webui-controlnet\annotator\downloads\openpose\hand_pose_model.pth

10%|███████▌ [

i use openpose full why he try download hand pose model?

[ornate elk]

Is the preprocessor set to openpose full?

[chilly knot]

yes

if use diffrent openpose model RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

problem only if i choise preprocessor

[vocal burrow]

try removing any non latin characters in the path (like the dot in sd.webui)

[chilly knot]

hm now open pose read pose from image and apply

also i have bad wifi

and often i have SSL error

when download someting in sd

[crystal nacelle]

Hey again. I’m still having the connection error. Updated to python3.10.11. Added antivirus exceptions and no adblocker. Anyone know whatsup?

I installed an antivirus package that had anti tracking and stuff recently, maybe it’s that?

[ornate elk]

That could block, but what the cmd shows?

[crystal nacelle]

Thanks for your time CS1o. my CMD shows "1006" when i try to generate image and get the connection error. am launching web UI now to show you the full cmd...sorry for long post everyone

D:\stable-diffusion-webui-directml>git pull
Already up to date.
venv "D:\stable-diffusion-webui-directml\venv\Scripts\Python.exe"
fatal: No names found, cannot describe anything.
Python 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]
Version: 1.8.0-RC
Commit hash: eb286516b35b6bd5ad12ff902292290ba7de89d2
no module 'xformers'. Processing without...
no module 'xformers'. Processing without...
No module 'xformers'. Proceeding without it.
D:\stable-diffusion-webui-directml\venv\lib\site-packages\pytorch_lightning\utilities\distributed.py:258: LightningDeprecationWarning: pytorch_lightning.utilities.distributed.rank_zero_only has been deprecated in v1.8.1 and will be removed in v2.0.0. You can import it from pytorch_lightning.utilities instead.
rank_zero_deprecation(
You are up to date with the most recent release.
Launching Web UI with arguments: --use-directml --upcast-sampling --opt-sub-quad-attention --opt-split-attention-v1 --no-half-vae --no-half --use-zluda --lowvram --update-check --skip-ort

===========================
You are running torch 2.0.0+cpu.
The program is tested to work with torch 2.1.2.
To reinstall the desired version, run with commandline flag --reinstall-torch.
Beware that this will cause a lot of large files to be downloaded, as well as
there are reports of issues with training tab on the latest version.

Use --skip-version-check commandline argument to disable this check.

Civitai Helper: Get Custom Model Folder
Tag Autocomplete: Could not locate model-keyword extension, Lora trigger word completion will be limited to those added through the extra networks menu.
[-] ADetailer initialized. version: 24.1.2, num models: 9
ControlNet preprocessor location: D:\stable-diffusion-webui-directml\extensions\sd-webui-controlnet\annotator\downloads
2024-03-27 21:59:28,622 - ControlNet - INFO - ControlNet v1.1.440
2024-03-27 21:59:28,808 - ControlNet - INFO - ControlNet v1.1.440
Loading weights [906f4a33c9] from D:\stable-diffusion-webui-directml\models\Stable-diffusion\comicBabes_v2.safetensors
2024-03-27 21:59:29,885 - AnimateDiff - INFO - Injecting LCM to UI.

2024-03-27 21:59:33,716 - AnimateDiff - INFO - Hacking i2i-batch.
2024-03-27 21:59:33,825 - ControlNet - INFO - ControlNet UI callback registered.
Creating model from config: D:\stable-diffusion-webui-directml\configs\v1-inference.yaml
Civitai Helper: Settings:
Civitai Helper: max_size_preview: True
Civitai Helper: skip_nsfw_preview: False
Civitai Helper: open_url_with_js: True
Civitai Helper: proxy:
Civitai Helper: use civitai api key: False
Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch().
Startup time: 154.4s (prepare environment: 45.4s, initialize shared: 6.9s, other imports: 2.3s, opts onchange: 125.9s, setup codeformer: 0.4s, setup gfpgan: 0.1s, list SD models: 0.3s, load scripts: 6.1s, initialize extra networks: 0.7s, create ui: 6.0s, gradio launch: 0.8s).
1006

oh shiz, am seeing now that maybe i should try reinstall-torch.

[vocal burrow]

no it s okay for directml fork to be stuck at 2.0 for now

although you can t have both --use-zluda and --use-directml working at the same time

[crystal nacelle]

okay let me remove one

[static panther]

I have downloaded and used A1111 on my desktop twice with little to no issues but now that i'm downloading it to my laptop it seems to have a lot of issue.
I'm currently stuck scrolling the internet trying to find the fix to this

[vocal burrow]

what s the full console log ?

[ornate elk]

There is something wrong here

What's your GPU?

[crystal nacelle]

AMD RX7600 XTX 8GB

[static panther]

the log said the same thing before I closed it

[ornate elk]

You followed my zluda guide I asume?

Because you still have commands from the old directml guide in there

[vocal burrow]

we need to see it tho. it displays a lot more helpful information than what is shown in your screenshot

[crystal nacelle]

I did. but Aryetis said i cant have --zluda and --usedirectml. i removed zluda. let me put it back and remove directml instead.

am using this command line now:
set COMMANDLINE_ARGS= --upcast-sampling --opt-sub-quad-attention --opt-split-attention-v1 --no-half-vae --no-half --medvram-sdxl --use-zluda --update-check --skip-ort --reinstall-torch

[static panther]

I closed the log to try to fix a few things, I reopened it and i'm stuck on the infinite checkpoint loading bug

[ornate elk]

If you followed my zluda guide you only should use:
--use-zluda --medvram-sdxl --skip-ort --update-check
Remove the rest

[vocal burrow]

then show us that log

[crystal nacelle]

okay. let me try that.

[ornate elk]

The guide is in the pinned message of this channel:
#🤝｜tech-support message

[static panther]

It's been stuck like this for the past few mins

[ornate elk]

What's your gpu?

[static panther]

NVIDIA GeForce RTX 3060

[ornate elk]

The 12gb vram version?

[static panther]

6gb dedicated memery
13bg GPU Memory

[ornate elk]

Okay then close the cmd.
Edit the webui-user.bat and there remove --skip-torch-cuda-test
Then delete the booru2prompt folder inside the Extensions folder.
Then relaunch the webui-user.bat

[crystal nacelle]

am getting the "torch not able to use GPU" error. so am going to add "--skip-torch-cuda-test", is that okay?

[static panther]

that's what i did

[ornate elk]

Okay now close the cmd.
Open up a cmd and type
Pip cache purge
Then hit enter. And close the cmd.
Then delete the venv folder.
Relaunch the webui-user.bat

[ornate elk]

make sure you followed every step mentioned in the guide. starting at "Getting zluda to run" step 1

[crystal nacelle]

okay, im sure i did but ill do it again.

[ornate elk]

and dont add --skip-torch-cuda-test

[static panther]

Curse my slow internet

[ornate elk]

rip :/ but then it will work

if the download and install doesnt get interrupted

[static panther]

if it does?

[ornate elk]

if it does you can relaunch the webui-user.bat and it tries to proceed where it left

and if you get errors you would need to delete the venv folder again

[static panther]

thxz

[ornate elk]

no problem, the issue you had was that with --skip-torch-cuda-test you have let it download the wrong files. so instead of using your GPU it has run on the CPU only

and now its downloading the gpu version of the file

[crystal nacelle]

this step confuses me "Move one 1.5 based model from your other existing directml webui install into the models/stable-diffusion folder. Or skip this if you dont have one."
my questions: how do i differentiate between a 1.5 model and a normal model?
Should i launch the new SD-Zluda webui?

[ornate elk]

yes launch it

a "normal" model is an 1.5 model

[crystal nacelle]

any of these 1.5?

maybe dreamshaper?

[ornate elk]

yes try dreamshaper

[pastel steeple]

Hi folks! Anyone having issues with Stability API? Everything worked fine for me 20 mins ago and now my requests just hang for ever. I am using Stable Image Core API

[crystal nacelle]

im downloading zluda from here
https://github.com/lshqqytiger/ZLUDA
instead of:
https://github.com/lshqqytiger/ZLUDA/releases

[ornate elk]

download the latest stable version of the releases

the ZLUDA-windows-amd64.zip

[crystal nacelle]

okay thank you. i almost nuked myself with the wrong download

[vocal burrow]

never use that "Download ZIP" option unless explicitly stated otherwise

[crystal nacelle]

AMD HIP SDK is installing all full components except Pro23.Q4 display driver, is that okay?

[crystal nacelle]

ah! i see, thank you Aryetis

[vocal burrow]

otherwise you ll download source code instead of actual executable files

[crystal nacelle]

like a noob

[ornate elk]

yes every hip sdk but not the pro driver

[quick widget]

Same for me 👍

[vocal burrow]

@pastel steeple @quick widget weird, everything seems fine on their side https://stabilityai.instatus.com/, maybe it has something to do with your ISP ?

[quick widget]

I don't think so because it does not work in local and on the cloud :/

[crystal nacelle]

im on the final step. i guess, from now one i will use this new SD-zluda. am i supposed to edit this or should i paste it as it is "--ckpt-dir "Path/to/models/stable-diffusion" --lora-dir "Path/to/models/lora" --controlnet-dir "..." --embeddings-dir "...""

[vocal burrow]

if it doesnt work in local then it has nothing to do with stability api ?

not sure I understood you on this one

[ornate elk]

you would need to edit this to match your paths

[quick widget]

I mean doing a curl call wether it's from my machine or from my cloud provider, the http calls get stuck forever

The server never seem to answer

[vocal burrow]

:/ maybe you can try to report the issue / get in contact with them through mail https://beta.dreamstudio.ai/support

[quick widget]

Already did thanks 🙏

[vocal burrow]

please let us know about their answer

[crystal nacelle]

awesome. thanks. for example, i would do this "--ckpt-dir "D:\stable-diffusion-webui-directml\models\Stable-diffusion" --lora-dir "D:\stable-diffusion-webui-directml\models\Lora" --controlnet-dir "D:\s...""

but id have to re-install extensions manually, no work around for that?

so from now on i should use this new sd-zluda/webui-user.bat right?

[vocal burrow]

you can symlink the extensions folder. not sure i d recommend it thought. It s probably for the best to not mix and match code between the two versions.

[ornate elk]

you can copy the extensions over into the extensions folder

but for heavy extensions a reinstall is better

[crystal nacelle]

copy it is XD

[pastel steeple]

Same here

[pastel steeple]

Looks like it's fixed

[vocal burrow]

:/ ok that s weird

oh yeah it s finally showing up on the status page https://stabilityai.instatus.com/

[pastel steeple]

Does this happen often? I'm planning to integrate Stable in my SaaS, don't want to get chased by users on a daily basis 😄

[vocal burrow]

well you can check the issues history on the page I ve linked

[pastel steeple]

under the assumption all downtimes are logged there 🙂

[vocal burrow]

I have no reasons to think otherwise but yes

[pastel steeple]

Awesome! We've been using Dall-E-3 thus far and I'm fed up with it, stable seems like a massive step up, at least at first sight

[crystal nacelle]

okay its working, still got the connection error on img2img. aswell as 2 warnings; WARNING: ZLUDA works best with SD.Next. Please consider migrating to SD.Next.
AND pytorch_lightning.utilities.distributed.rank_zero_only has been deprecated in v1.8.1 and will be removed in v2.0.0. You can import it from pytorch_lightning.utilities instead.

[feral parrot]

Also experiencing this outtage. getting a 504 error. For API calls I've found stability to a little less reliable thus far compared to other providers. Make sure to test a lot before switching over

[vocal burrow]

what is your gpu ?

[crystal nacelle]

AMD RX7600 XTX 8GB

[vocal burrow]

and what command line argument are you using atm ?

[pastel steeple]

Yeah, the status page shows a couple of pretty long downtimes but I hope it's due to the API still being a new thing for the company. Fingers crossed they keep making it more Stable 🙂

[crystal nacelle]

--use-zluda --medvram-sdxl --update-check --skip-ort

--ckpt-dir "D:\stable-diffusion-webui-directml\models\Stable-diffusion" --lora-dir "D:\stable-diffusion-webui-directml\models\Lora" --controlnet-dir "D:\stable-diffusion-webui-directml\models\ControlNet" --embeddings-dir "D:\stable-diffusion-webui-directml\embeddings"

[vocal burrow]

xd nice name, anyways for img2img to work on AMD I think you still need to use --no-half. I don't have an amd gpu to confirm tho. (cs1o will confirm once they get back as they have an amd gpu)

[quick widget]

Api working again thanks!

[vocal burrow]

and about pytorch_lightning.utilities.distributed.rank_zero_only we d probably need a full console log to understand what s going on

[feral parrot]

Have you tried replicate for sdxl it's reliable, just doesn't have this new 'core' version which is nice

[chilly knot]

i have download torch error help how i need instal it manual

cp310-win_amd64.whl

i need venv activate script

go to sd folder and install?

[pastel steeple]

I haven't, I'm still pretty new to Stable... I saw the Stable Image Core being marketed as their top of the line model and went straight for that. I'll check SDXL as well, thanks!

[ornate elk]

that can be ignored

i get these too

[ornate elk]

nope, --no-half is not needed for zluda

[ornate elk]

for the connection error, please whitelist the webui in any browser adblocker

[crystal nacelle]

other than adblocking whitelisting it, what other problems could be causing the connection error?

[ornate elk]

VPN, Proxys, Pi-hole, other adblockers

Third Party Antivirus

script blocker browser extensions

[crystal nacelle]

okay let me look around. i suspect the new antivirus pack i installed is causing this.

[ornate elk]

whats your antivirus? (98% of third party antivirus are bad and bloated)

[crystal nacelle]

AVG

[ornate elk]

thats one you dont need

AVG got bought off by a shady company called NortonLifeLock360

[crystal nacelle]

I remember Norton Antivirus, lol. okay i unistalled it and sure enough, connetion error isnt there anymore. im seeing this error when i try generate ZludaStable:

return torch._C._cuda_memoryStats(device)
RuntimeError: invalid argument to memory_allocated

So i added the --medvram argument and still getting same error

[ornate elk]

dont just add commands

the error is because you probably need to redo the step with the rocm librarys folder

or does txt2img work?

[crystal nacelle]

let me try it now. i think the connection error is legit gone tho, thank god. which step is the rocm librarys folder step?

[ornate elk]

first try to generate an image in txt2img

or do you get the error there?

the rocm librarys folder is step 4

[crystal nacelle]

same error

File "D:\SD-Zluda\stable-diffusion-webui-directml\venv\lib\site-packages\torch\cuda\memory.py", line 309, in reset_peak_memory_stats
return torch._C._cuda_resetPeakMemoryStats(device)
RuntimeError: invalid argument to reset_peak_memory_stats

With txt2img

[ornate elk]

okay then recheck step 4 of the zluda part

[crystal nacelle]

4. If your GPU is below a RX 6800?

[ornate elk]

let me recheck

[crystal nacelle]

but no i have....im running 76...i havextx76zerzero...man this fukn card keeps disappointing me. cvouldnt mine crypto with it either ima sell it for an nvidia asap

[ornate elk]

yes your card is newer, but we need to check if rocm supports it

[ornate elk]

okay, i rechecked and you dont have to do the library step

@crystal nacelle you need to do the following:
go into the C:\ZLUDA\ Folder.
There copy the cublas.dll and the cusparse.dll and paste them inside it again.
Rename the copys to cublas64_11.dll and cusparse64_11.dll
Copy these two files to stable-diffusion-webui\venv\Lib\site-packages\torch\lib, and overwrite if asked.

[crystal nacelle]

okay let me try that

No luck, still getting the same error.

[ornate elk]

can you show me the versions at the bottom of your webui?

it should look like this:

[chilly knot]

Installing collected packages: mpmath, typing-extensions, sympy, networkx, MarkupSafe, filelock, jinja2, torch
ERROR: Exception:
Traceback (most recent call last):
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\pip_internal\cli\base_command.py", line 180, in exc_logging_wrapper
status = run_func(*args)
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\pip_internal\cli\req_command.py", line 245, in wrapper
return func(self, options, args)
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\pip_internal\commands\install.py", line 452, in run
installed = install_given_reqs(
File ":\Stable diffusion\stable-diffusion-webui\venv\lib\site-packages\pip_internal\req_init_.py", line 72, in install_given_reqs

when i intall torch manual

[haughty scarab]

Hello

I am new to sd but it seems like it cant load anything

It doesnt load vae, checkpoint, lora.....

And it doesn't give an error

[ornate elk]

hey, what does the cmd shows?

[haughty scarab]

Wait

I am duplicating it

[chilly knot]

Error -3 while decompressing data: invalid block type when i install whl

its bad file?

[haughty scarab]

Cmd and waht it looks like

[haughty scarab]

?

[ornate elk]

what webui do you use?

and where do you use it?

[haughty scarab]

Forge

[haughty scarab]

What do you mean

[ornate elk]

its a local install right?

[haughty scarab]

It was the same with automatic 1111

[haughty scarab]

Yeah

[ornate elk]

whats your GPU?

[haughty scarab]

Geforce nvidia 940m

[chilly knot]

where you live? time different on 30 min

with me

[haughty scarab]

Iran

[chilly knot]

yes time correct

[ornate elk]

make sure you whitelist the webui in any browser adblocker

also try and other browser if it still not work

[haughty scarab]

How?

[chilly knot]

cs10 can you answer : invalid block type when i install whl
its bad file?

[ornate elk]

you shouldnt install the whl file manually

when deleting the venv, the right files already gets installed

or what do you want to try?

[ornate elk]

first try opening Edge browser and open the same url there and test

[chilly knot]

i cant download pytorch ssl error mac beacouse bad wifi

with bat file

i go to venv activate

and send pip instal torch whl

[chilly knot]

sd and python is hell)

[chilly knot]

cs1go python change files if they bad ?or only check names in folders

[ornate elk]

python?

[chilly knot]

when you start sd

from bat

[ornate elk]

it tries to load the venv

[chilly knot]

how i understand that my torch files ok

[ornate elk]

if you dont get errors

make sure you have python 3.10.11 installed

[chilly knot]

i have it when i start canny

and preprocessor

[ornate elk]

can you show a screenshot of you controlnet settings?

also how big is the canny model file?

[chilly knot]

1.4 gb

and i have one canny 5 gb

both give me error

[ornate elk]

okay for SD 1.5 please use these controlnet files, here is canny:
https://civitai.com/models/38784?modelVersionId=44716

[chilly knot]

its nor for sd 1.5?

[ornate elk]

its for 1.5

its just version 1.1 of controlnet models

[chilly knot]

my files

[ornate elk]

probably

[chilly knot]

big size?

[ornate elk]

for your gpu the big size controlnet models are bad

the one i linked you is only 700mb

and works fine for everyone

[chilly knot]

after i delete venv new folder more on 200 mb now and installing something

can you tell me how size your torh folder

[ornate elk]

venv is 6.4gb for me

[chilly knot]

torch folder in lib

[ornate elk]

4.5gb

[chilly knot]

4.57 my

[cerulean tusk]

It seems the search and replace API isn't working anymore. The colab example for it stopped working as well. Anyone know anything about this?

[next oak]

Hi guys, I'm a begginer here. I finalized an image but I want to add two details on it. Do you know how can I do to add them and keep the image as it is ?

[vocal burrow]

They've (stability) had issues with the API earlier today, maybe they're still fixing some stuff. cf https://stabilityai.instatus.com/

[olive patio]

hello all! im having a hard time stomping a bug. im run stable video diffusion locally and for some reason it works with some images and not others. i get an unboundlocal error about the input image. I've checked everything i can find and can't figure out why this image won't work. it was suggested that I update, but i only downloaded everything a few days ago.

UnboundLocalError: cannot access local variable 'input_image' where it is not associated with a value
this happens when i try and use simple_video_sample.py

[austere charm]

I used to be able to use Lora's on the "additional network" tab a while back, but after some updates it no longer loaded from there so I switched to putting Lora's into the text prompt, I was curious if there was a way to gain the additional network Lora function back? Currently on 1.8.0

[ornate elk]

nope its outdated and doesnt work anymore

currently no other solution then adding them at the lora tab

[austere charm]

Ah for sure, thanks for the info! are there any suggestions of recreating images found on civitai then? I can see loras someone used via the "Resources Used" section, but some of them don't have weights in the text prompt

[olive patio]

I figured out the problem. there is a bug in the simple_video_sample.py script. if the input image is RGB mode than the variable input_image never gets set and results in the error. i fixed it with an else statement that catchs the case when the image.mode starts as RGB

[valid pewter]

can anyone help me, im trying to use instant id, and whenever i do this, i get outcomes like this.

[vocal relic]

Anyone got a good inpainting method? I mainly do one shots and find inpainting rarely gives good convincing results for me.

[severe kindle]

The main issue I had with it is, that it needs a really small CFG, try 1.
Although I didn't get things quite that bad.
I usually pair it with a model that needs a low CFG too, like a LCM model, or model with the LCM LoRA

[valid pewter]

oh ok ill try

[crystal nacelle]

im running cu121

[ornate elk]

Yes thats a problem, (as its not supported by zluda).
open up a cmd and type
Pip cache purge
And hit enter.
Then delete the venv folder and relaunch the webui-user.bat

That should download the right cu version

[crystal nacelle]

okay thank you. let me try this next

[crystal nacelle]

okay i think its working. ive left it doing that long initial generation that could take 20min or so. god i hope it works

[crystal nacelle]

the pinned github guide has "Compilation, Settings, and First Generation" as a final setup but i cant find any of these parameters in my webui settings. are they a must-do?

[ornate elk]

You dont need them

[ornate elk]

Let me know if it worked

[crystal nacelle]

it worked bro! thank you so much

[ornate elk]

Perfect no Problem 🙂

[noble belfry]

Just come across a weird quirk. a default setting is save large images as jpg if they are oversized. This saves the parameters as an exif comment too which means you can use windows search to look for images by prompt etc. from explorer

[balmy lily]

i followed ur tutorial and i got this error when i ran the bat script

[ornate elk]

How much RAM do you have?

Error says no memory and some connection errors

[balmy lily]

i hav a 4060 ti, it said on the task manager that my gpu has 8.00 gb of dedicated memory
installed ram
16.0 GB (15.9 GB usable)

[balmy lily]

i tried running the bat script with every --vram command

[ornate elk]

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

16gb Ram should work

[balmy lily]

wait so is 4060ti not strong enough, what do u hav

[ornate elk]

Its enough

--xformers --medvram-sdxl --no-half-vae is needed in the webui-user.bat

But the problem here has nothing to do with the GPU

[balmy lily]

how can i increase my vram without changing my gpu

[ornate elk]

You dont need to

Like I said the problem is neither your vram or your GPU

A 4060ti is capable of running SD without problems

[balmy lily]

oh lol i misread one of ur messages, for some reason i thought you said

8 gb of RAM could be a problem

but it doesnt look like u said anything similar to that so idk how i read that

[ornate elk]

Yea 8gb RAM would be a problem but you said you have 16gb

What you should try is to delete the venv folder and relaunch the webui-user.bat

[balmy lily]

i have 16gb ram and 8 vram, assuming deciated gpu memory is vram

[balmy lily]

didnt work

[ornate elk]

Same error?

[ornate elk]

How much free space on C drive?

[balmy lily]

23gb

my c drive is alot smaller than my other drives, its only 100gb total, whereas my d and e are 400

[ornate elk]

Okay, open up a cmd
And run
Pip cache purge
Then delete the venv folder again

Then relaunch the webui-user.bat

[umbral fox]

Hello, I'm still not that far into Stable Diffusion yet. I am not that familiar with Stable Diffusion yet, so I need your help.

I used to have a LoRA called "NSFW POV All In One SDXL".
https://civitai.com/models/144203/nsfw-pov-all-in-one-sdxl-realisticanimewd14-74mb-version-available

Today I installed another LoRA ("Anime Lineart / Manga-like" https://civitai.com/models/16014/anime-lineart-style) and the above "NSFW POV All In One SDXL" is not displayed on the Stable Diffusion web UI.

[error description].
LoRA entry on Stable Diffusion.
Nothing here, please add content to the following directory
C:\Users\xxxxxxx\stable-diffusion\stable-diffusion-webui\models\Lora
C:\Users\xxxxxxxxx\stable-diffusion\stable-diffusion-webui\models\LyCORIS

[Current environment]
Stable Diffusion version:1.8.0

[What we did]
Place NSFW POV All In One SDXL data in the appropriate directory according to the error message.
Reload the UI.
Reboot PC
Upgrade Stable Diffusion (from 1.7.0 to 1.8.0)
Deletion of Anime Lineart / Manga-like
Reinstalled NSFW POV All In One SDXL
Update LoRA
From a different environment (I have another PC and am creating the exact same environment),
　Copy and paste the LoRA folder from another environment (I have another PC and am creating the exact same environment).

In all of these cases, it still does not show up.

What is strange is that on the other PC, which has the exact same configuration
NSFW POV All In One SDXL" and "Anime Lineart / Manga-like" can coexist.

Is there any good way to deal with this?
Please let me know if there is any other information necessary for making a decision.

[balmy lily]

will this be good for the bat script set COMMANDLINE_ARGS= --xformers --medvram-sdxl --no-half-vae

whats sdxl

also it seems to be working, but i changed --medvram-sdxl to --medvram, becuz --medvram-sdxl is what i did the first time

nvm it didnt work

[ornate elk]

The NSFW POV Lora is for SDXL models.
The Anime Lineart Lora is made for 1.5 based models.
The loras only show up when a compatible model is selected

[ornate elk]

Change it back to --medvram-sdxl

Its for the best performance for 8gb gpus

Also the error is very strange.
Can you restart your PC? Delete the venv folder and relaunch the webui-user.bat?

[crystal nacelle]

@ornate elk holy shiz dude Zluda is a game changer. I can see why you were loosing your shit about it…Generating content in seconds… huge win for the AMD squad, thanks again for helping me getting it to work!

[ornate elk]

Yea its just awesome! And no Problem.
Also if you try sdxl models and you run out of vram, try activate the FP8 mode for sdxl in the Settings

That reduces the vram usage of sdxl models

[balmy lily]

do i put the models in before or after running the bat script. also do u know how to install comfyui instead

[umbral fox]

So, it would be more unusual for it to be displayed all the time. ......
On the other hand, do you know why it is always displayed?

[ornate elk]

You should have 1 model inside the models/Stable-diffusion folder

[balmy lily]

are u sure i should do this installation

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

u say in ur nvidia tutorial
instead of this one from ur other tutorial

git clone https://github.com/lshqqytiger/stable-diffusion-webui-directml && cd stable-diffusion-webui-directml && git submodule init && git submodule update

[ornate elk]

Yes, you need to switch the checkpoint and then press the reload button at the lora tab

[balmy lily]

before running the web ui bat?

[ornate elk]

Yes I'm sure. As one guide is for nvidia GPUs und one guide for AMD

[ornate elk]

It could help yes

[balmy lily]

oh lol i didnt notice amd, i thought it was for just non nvidia or not necessarily nvidia, like i thought the nvidia one was just a slightly simpler optional way to do it if u happen to have an nvidia gpu

[ornate elk]

Not the case for SD. Nvidia and AMD need a different setup

@balmy lily to get back to your problem.
Are you connected over WiFi or Ethernet?
Are you using a VPN or not?
Located in Asia/Russia or not?

[balmy lily]

im on nordvpn, and im on wifi

[ornate elk]

Ah okay

The VPN could be an issue then
It seems the download of pytorch fails

[umbral fox]

I learned a lot! Thank you!

[ornate elk]

No problem 🙂

@balmy lily if you can try disable the VPN, then delete the venv folder and relaunch the webui-user.bat

[balmy lily]

what should the model be and what exact folder should it be in

[ornate elk]

To test you can download the Dreamshaper v8 model from here and place it inside the stable-diffusion-webui/models/Stable-diffusion folder
https://civitai.com/models/4384/dreamshaper