Entry Point Not Found

[Desuka-art]

Is there an existing issue for this?

• [X] I have searched the existing issues and checked the recent builds/commits

What happened?

After installing xformers using a rentry guide, suddenly I get two errors, one as a popup, and another on the command line.

Please help!

Steps to reproduce the problem

Hard to say, I did a lot in order to get xformers to be recognized, so... I don't know.

What should have happened?

The UI should have launched normally, with xformers.

Commit where the problem happens

Launching Web UI with arguments: [WinError 127] The specified procedure could not be found WARNING: [WinError 127] The specified procedure could not be found Need to compile C++ extensions to get sparse attention support. Please run python setup.py build develop

What platforms do you use to access UI ?

Windows

What browsers do you use to access the UI ?

No response

Command Line Arguments

No response

Additional information, context and logs

No response

[CJohnDesign]

I'm running into the same error on TheLastBen colab

[StiffPvtParts]

Having the same issue in colab.

EDIT: @CJohnDesign Try deleting (or renaming to be safe) the SD folder in your Google Drive and then running all the batches of code in the Colab notebook to set up the folder again. Worked for me a moment ago.

[EvaSamuelCodes]

I'm having the same issue. Just installed xformers. Happens when I run launch.py locally on my windows 10 machine

[albertotrunk]

windows 11 dev

[asimard1]

I tried reinstalling xformers using the .whl file, and the --force-reinstall argument... This completely broke my installation of the UI and now I get this error whenever I try generating images, even without xformers. error.txt

The first error I can see appear is

A matching Triton is not available, some optimizations will not be enabled.
Error caught was: No module named 'triton'

So don't try doing that... I'll try solving by myself but if anyone has a solution it would be appreciated!

Edit: Uninstalling xformers from venv solves the problems for now, but obviously doesn't help for the installation of xformers. At least the UI can run without it.

[kabachuha]

@asimard1 Try installing the nighty build for triton. That fixed your error message for me

pip install triton==2.0.0.dev20221120

[aliencaocao]

If xformers is not compiled against the exact same version of Pytorch you are running, you have to recompile it or install the correct pytorch version.

[rasamaya]

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

[acoolrocket]

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

This worked for me too, automatically went to installing pytorch without faffing with gitbash/python cmd. Also make sure to keep it in the webui-user.bat after installation because it'll just toss up the error again.

Just gained almost 2x speed on my RTX 3060 laptop, feels great.

[illtellyoulater]

I got a very similar error message after renaming config.json from the Web-UI root dir to config.json.bak as I was trying to force Web-UI to write a new default config file... From that specific moment on every successive boot I would get the "Entry Point Not Found" error,

In the end, renaming "config.json.bak" to "config.json" reverted the situation to normal, without no further errorsa...

[matichek]

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

Yes, perfect, this worked.

[dtrevithick]

This worked great for me. Anyone know if its reinstalling or adding a new torch and pytorch download? Dont want to have multiple installs of this.

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

[MSire]

Tried it but does not work for me. Here is the issue I get : Unable to create venv in directory ""C:\Ai\stable-diffusion-webui\venv\pyvenv.cfg"" any idea ?

[shancf]

Tried it but does not work for me. Here is the issue I get : Unable to create venv in directory ""C:\Ai\stable-diffusion-webui\venv\pyvenv.cfg"" any idea ?

please replace pyvenv.cfg with pyenv.cfg

[MSire]

Just tried it but I now get another error when runing my webui-user.bat

[CaptainKokomo]

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

For those people who are copying and pasting this exact line.... This is an example of the what line is needed. Copy the line of the location of the file in the error message you are receiving.

[BlueLovagSD]

SOLVED

Yesterday I suddenly ran into the NaN error problem. I have a pull request in my webui-user.bat and I think it has something to do with pulling something broken. I then installed X-Formers and had the entry point problem:

adding a line to my webui-user.bat for example... set VENV_DIR=E:\Projects\AI_art\Automatic1111\stable-diffusion-webui\venv\pyenv.cfg Fixed for me, it reinstalled pytorch again, so look for folder venv in your install directory.

This worked for me but not completley. I can only create text2img, controlnet tiles doesn't work. After also activacting full precision and no half, so basically trying out all arguments, controlnet img2img with tiles don't work. When using x-formers it just freezes and without it just gets a NaN error.

Now I wanted to get rid of x-formers etc. I see this:

I updated PyTorch 2 times yesterday, it doesn't seem to work with pip somehow, even when it says it installed it. Is there a way to maybe fix the PyTorch issue first without pip? I'm just a UX-designer and i'm very unfamiliar with python.

When I want to update PyTorch now it says this:

Upon further trying things out I did: pip3 install numpy --pre torch --force-reinstall --index-url https://download.pytorch.org/whl/nightly/cu117

apparently I need cuda 18 so: pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

jesus christ... the download from pytorch says it's version 2.0.1

After starting webui again, I saw it still doesn't recognize any of the updates. I updated python to 3.10.11 and I DELETED the venv folder. Upon starting webui again, it downloaded the neccessary stuff automatically. After using --xformers it installed xformers again and now it seems to be working!