Skip to content

Machine-learning image upscaler with GUI; all dependencies integrated, no set-up

License

Notifications You must be signed in to change notification settings

jkendall327/waifu2x-ncnn-vulkan-gui

Repository files navigation

waifu-2x-nccn-vulkan-gui

build and test

Screenshot of the program

  • GUI frontend for waifu-2x-nccn-vulkan, created by nihui
  • Batch processing available
  • No dependencies, download and run

You can use this to upscale images (particularly drawings, rather than photographs).

nccn-vulkan is self-contained, so you should't need to install any other dependencies to get it working.

This is a thin wrapper around the console app, so it provides a live preview of what command is going to be run.

Select multiple images to process them all, one at a time.

Installation

The Releases tab contains a .deb installer which you can download and run to install the app.

Once it's installed, run waifu2x-gui in a terminal to launch the app.

Currently I've only compiled this for Linux, but it would probably work alright on Windows or OSX with minor tweaks.

Building

This project was built with .NET 6 and Avalonia.

If you have both of those installed, there's nothing weird with the build so you should be able to just clone and build the solution.

Step-by-step:

  1. Download and install .NET. You will need a developer SDK, not just the runtime.
  2. Ensure .NET is in your path. You can run dotnet --list-sdks to see if it works.
  3. Clone the repo.
  4. Navigate to the repo on your hard-drive.
  5. Run dotnet run (to build the app and start it) or dotnet build to compile an executable.

Known issues

Due to a bug in how Avalonia handles non-Latin characters on Linux, the app may crash on startup with a message about not being able to load a default font-family.

If this happens, launch the app with a modified locale environment variable:

LC_ALL=C [executable path here]

If you're building from source, you can also try this:

LANG=en_US.UTF-8 dotnet run

Contributions

Highly welcome!

About

Machine-learning image upscaler with GUI; all dependencies integrated, no set-up

Resources

License

Stars

Watchers

Forks

Packages

No packages published