Being a Microsoft Store app carries too much unexpected issues, while I have no intention of make a Microsoft Store app here,

This is a really cool program you've built, and I prefer it over easy diffusion or others. I also appreciate the built-in NSFW filter. Some things are better not seen. However, I had one of my prompts flagged as an explicit image.
Positive:
Jagged snow covered peaks rise from a snow covered forest, above a lake. A massive glacier descends down the mountain into the lake. realistic, crisp, Alps, Cascades
negative:
blurry, render, fuzzy

model: realistic vision onnx

Settings:
512X512
batching: 9
steps: 75
random seed: 1391526699
guidance strength: 7

The Assets won't download, the error Git gives is

Downloading Unpaint/Assets/LockScreenLogo.scale-200.png (4.1 KB)
Error downloading object: Unpaint/Assets/LockScreenLogo.scale-200.png (7fd70ec): Smudge error: Error downloading Unpaint/Assets/LockScreenLogo.scale-200.png (7fd70ec6e4e2854c4e67c15b939b21a04c90e4a4712a20eb530866499792a9b8): batch response: This repository is over its data quota. Account responsible for LFS bandwidth should purchase more data packs to restore access.

Errors logged to 'C:\Cpp\unpaint\.git\lfs\logs\20230609T123116.492325.log'.
Use `git lfs logs last` to view the log.
error: external filter 'git-lfs filter-process' failed
fatal: Unpaint/Assets/LockScreenLogo.scale-200.png: smudge filter lfs failed

They aren't in the source zips either, including the one with the latest release.

The solutionn I found for the time being was to grab the MSIX and extract the Assets from it with 7zip. Maybe this will be helpful to others.

!_onnxEnvironment expression should be evaluated before usage of _onnxEnvironment->DeviceId in the code void StableDiffusionModelExecutor::EnsureEnvironment(std::string_view modelId) { if (_modelId != modelId || _onnxEnvironment->DeviceId != int32_t(_unpaintOptions->AdapterIndex()) || !_onnxEnvironment)

On Windows 11, app header works glitchy. I'm using the Light mode, and Min/Max/Close buttons are not visible without mouse hovering:

Also, the whole app content (UI, image) randomly (not often) blinks.

when i follow the guide of Model-import, then run "python .\stable_diffusion.py --optimize --model_id runwayml/stable-diffusion-v1-5" in "olive-sd\olive\examples\directml\stable_diffusion" directory, report error as follow:

(olive-env) PS D:\E_2_D\temp\axodox\olive-sd\olive\examples\directml\stable_diffusion> python .\stable_diffusion.py --optimize --model_id runwayml/stable-diffusion-v1-5
C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\pydantic_internal_config.py:269: UserWarning: Valid config keys have changed in V2:

• 'json_dumps' has been removed
• 'json_loads' has been removed
  warnings.warn(message, UserWarning)
  C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\pydantic_internal_config.py:269: UserWarning: Valid config keys have changed in V2:
• 'json_dumps' has been removed
• 'json_loads' has been removed
  warnings.warn(message, UserWarning)
  C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\pydantic_internal_config.py:269: UserWarning: Valid config keys have changed in V2:
• 'json_dumps' has been removed
• 'json_loads' has been removed
  warnings.warn(message, UserWarning)
  Traceback (most recent call last):
  File ".\stable_diffusion.py", line 18, in
  from olive.workflows import run as olive_run
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\olive\workflows_init_.py", line 5, in
  from olive.workflows.run.run import run
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\olive\workflows\run\run.py", line 16, in
  from olive.passes import Pass
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\olive\passes_init_.py", line 5, in
  from olive.passes.olive_pass import FullPassConfig, Pass
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\olive\passes\olive_pass.py", line 13, in
  from olive.common.config_utils import ConfigBase, validate_config
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\olive\common\config_utils.py", line 118, in
  class ConfigListBase(ConfigBase):
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\pydantic_internal_model_construction.py", line 98, in new
  private_attributes = inspect_namespace(
  File "C:\Users\synov\miniconda3\envs\olive-env\lib\site-packages\pydantic_internal_model_construction.py", line 291, in inspect_namespace
  raise TypeError("To define root models, use pydantic.RootModel rather than a field called 'root'")
  TypeError: To define root models, use pydantic.RootModel rather than a field called 'root'
  (olive-env) PS D:\E_2_D\temp\axodox\olive-sd\olive\examples\directml\stable_diffusion>

Hi, I have used optimum from huggingface to convert the model to onnx and this is impossible to open the model in unpaint because it doesn't contain safety_checker folder.

I have tried to change the code to don't use safety checker but I have one error "Saving initialized tensors (failed:open file model.onnx_data)"....
So, impossible to test your software actualy.
If you have a valid solution ?
Thanks.

Not a critial request, but it would be nice to automatically know that some new stuff is available. Thank you!

I have an integrated AMD and discrete Nvidia graphics (laptop). By default, Unpaint uses AMD, and it don't "see" Nvidia at all.
However, I was able to force it in Windows Graphics settings, and after that, it worked.
P.S. It could be useful to show some system info in Settings section or something; like GPU detected, RAM/VRAM used, etc.

Will be here the support for ROCm?

There is already c++ Stable Diffusion using ROCm, but with no UI, which is suck: https://github.com/leejet/stable-diffusion.cpp

It compiles fine, but does not generate anything (not sure if it is meant to).
I also wonder where I have to put the model files exactly for that (libfolder)...
What I would like to do: Run stable diffusion with Open Frameworks (maybe as an addon - there is already one for .onnx files).
For that I do not need Axodox::Graphics, because I would copy the pixels of the resulting tensor to an openGL texture...

See screenshot here.

I'm on Windows 10 21H2,

I'm not sure where the app was installed after running msix.
Where is the config file or some other way to specify existing models folder?

I would like to thank you for bringing such functionality to UWP application,
it works very well (Using it on my laptop), at least I don't have to run local server to use it.

I do have a little advice if you don't mind regarding the target users:
Microsoft already made big damage to UWP using unnecessary restrictions and wrong environment concept along strict Store policy.
the UWP apps mostly now used in XBOX,
that's why I'm not expecting to see much contributors (I hope I'm wrong about that),
UWP's first impression wasn't very good for developers so they still not very convinced to engage with it,
Windows users still prefer Win32 apps style over UWP because of the limitation,
unless they're looking for simplicity which you already achieved it in your app,
don't depend on UWP security market, users will trade security for features,
so I do suggest to checkout this discord server, and see if there is opportunity to post it in the announcement so it can get more attention from the right community, also users in this server already has experience to deal with dev mode.

I hope to see this app more popular and has more support.

Most things I liked:

• Results pretty good
• Safety check is really good
• Using Modify (option) instead of img2img as text was perfect (specifically for basic users)
• Mask tool seems easy to use unlike other apps.
• Super easy to use and modify generated images

What I hope to see:

• Inverted mask, so I can choose to keep one object and remove every thing else.
• Choose external location for output images (this maybe for XBOX only, desktop is fine)

btw I cannot use any higher resolution other than 512x512 with my NVIDIA RTX-2070 , Windows 10
I thought it's model issue but for some reason I can generate higher resolutions with other python apps.
If this is limited so I prefer to make 512 default on PC too, so it doesn't fail from the first time.
user will have to switch between models to generate again, otherwise it will stuck at failed state (happens with 768x768).

Sorry if that too long.

Many thanks, I really enjoyed using this app.

I installed
Unpaint_1.0.69.0_x64.msix
With dependency
Microsoft.UI.Xaml.2.8_8.230

Using the model
axodoxian/stable_diffusion_onnx

On my Xbox series S and I'm getting "Generation failed"

I tried the default (SraKmd_arden) and the Microsoft basic render driver.

I'm not sure what I'm doing wrong or how to troubleshoot the logs. It would be nice to include a more descriptive error message. Any guidance?

...so user can load existing models instead of copying them to the deafult directory on C:.
Thank you!

Any plans to integrate this? https://github.com/dinglufe/segment-anything-cpp-wrapper

Please consider adding support for Linux, as it can be beneficial, especially for deployment on cloud platforms.
Are there any limitations that prevent the code from running on Linux, or is it primarily related to the building process?
It appears that the code utilizes ONNX, which is supported on Linux.

any ways to disable the safety check? cant seem to find how (yes im shameless)

Hi, I updated the app to the latest version (43), and it stopped to work after updating. After full reinstalling, it worked untill I hit the Install additional tools in Settings.
Getting this error in logs:

AppName Unpaint.exe 
  AppVersion 0.0.0.0 
  AppTimeStamp 64aafa1f 
  ModuleName Windows.UI.Xaml.dll 
  ModuleVersion 10.0.22621.1928 
  ModuleTimeStamp e91f7675 
  ExceptionCode c000027b 
  FaultingOffset 0000000000872360 
  ProcessId 0x2228 
  ProcessCreationTime 0x1d9b3662506736d 
  AppPath C:\Program Files\WindowsApps\Unpaint_1.0.43.0_x64__asy5p9gk2jzay\Unpaint.exe 
  ModulePath C:\Windows\System32\Windows.UI.Xaml.dll 
  IntegratorReportId 4b40cd91-9ed6-444d-8435-5cce5c820e92 
  PackageFullName Unpaint_1.0.43.0_x64__asy5p9gk2jzay 
  PackageRelativeAppId App

Tried every model under the sun that was stable diffusion ONNX format but kept saying "The model does not match the Stable Diffusion ONNX schema." Tried to import locally said the same thing seems like a possible platform issue then a software issue because of the difference between windows and Xbox windows. But I'm not a expert on that since MS isn't really opened on that but hopefully this is a simple fix either on my part or yours. Or just might be my mistake.

Instead of another downloading, people who already have some bunch of models would be glad to import them, converting from safetensors/ckpt/diffusers to onnx.

1. How to disable NSFW checker and Safe Mode? There are no switchers in UI and Settings:
   
2. Got an error a couple of times, but have no idea which one and how to "debug". How to check error log and generating stats (speed, etc.)?
3. I see some hotkeys are working here (like Del to delete the image). How to see all of them? It would be nice to have some listing visible.

Thank you!

How to stop/pause the generation process if it stucks, etc.?

Using the latest version 1.0.10.0, I can't get the masking tool options to show on the left side of the app

Because it always trigger the exception below, for debugging the issue, I removed the code below:

    catch (...)
    {
      async.update_state(1.f, "Generation failed.");
      return {};
    }

Then I got the below exception call stack. Any idea on how to debug this issue? Thanks in advance!
I downloaded the models from 'https://huggingface.co/axodoxian/stable_diffusion_onnx':

>	Unpaint.exe!winrt::terminate() Line 5206	C++
 	Unpaint.exe!std::coroutine_traits<winrt::fire_and_forget,winrt::Unpaint::implementation::InferenceViewModel &>::promise_type::unhandled_exception() Line 9623	C++
 	Unpaint.exe!`winrt::Unpaint::implementation::InferenceViewModel::GenerateImage$_ResumeCoro$1'::`1'::catch$46() Line 386	C++
 	[External Code]	
 	Unpaint.exe!winrt::Unpaint::implementation::InferenceViewModel::GenerateImage() Line 332	C++
 	[External Code]	
 	Unpaint.exe!winrt::impl::resume_background_callback(void * __formal, void * context) Line 8915	C++
 	[External Code]	

Hi, do you consider to support of SDXL in some near future? Thanks!

Hi, I downloaded a canny model. Could you please give some giude, how to use it in real world? In UI I got this:

Thanks!

How to show this ControlNet diag log UI event putted downloaded "controlnet_onnx" folder under loaded "stable_diffusion_onnx" model by different ways (Unpaint store version 1.0.37.0)
https://huggingface.co/axodoxian/stable_diffusion_onnx/tree/main
https://huggingface.co/axodoxian/controlnet_onnx/tree/main

Firstly, congratulations on developing such an impressive project!

I am currently attempting to build this project using Visual Studio 2022, but I am encountering a specific error. Could you possibly provide some insight or guidance on why this might be happening? Any assistance would be greatly appreciated."

Follow the steps and installed all required package.
Updated all nuget packages to latest.
Windows 11
Visual Studio 2022 17.6.4

errors:

1>pch.cpp
1>C:\workspace\github\unpaint\packages\Axodox.Common.1.0.23\include\Infrastructure\Logger.h(27,48): error C2039: 'format_string': is not a member of 'std'
1>C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Tools\MSVC\14.33.31629\include\numeric(23): message : see declaration of 'std'

Hi, thank you for the great project.

After I hit Browse, the app suddenly closes. Is there any dump file I can send you or any static path I can put my models in?