nagadomi/nunif

My playground.

For the time being, I will make incompatible changes.

waifu2x

waifu2x/README.md

waifu2x: Image Super-Resolution for Anime-Style Art. Also it supports photo models (GAN based models)

The repository contains waifu2x pytorch implementation and pretrained models, started with porting the original waifu2x.

The demo application can be found at

• https://waifu2x.udp.jp/ (Cloud version)
• https://unlimited.waifu2x.net/ (In-Browser version).

iw3

iw3/README.md

I want to watch any 2D video as 3D video on my VR device, so I developed this very personal tool.

iw3 provides the ability to convert any 2D image/video into side-by-side 3D image/video.

iw3-desktop

iw3/docs/desktop.md

iw3.desktop is a tool that converts your PC desktop screen into 3D and streaming over WiFi.

You can watch any image and video/live displayed on your PC as 3D in realtime.

iw3-player

iw3/player/README.md

iw3-player is a self-hosted, specialized viewing environment for stereoscopic media.
It allows you to stream media that has been pre-converted to 3D with iw3 from your PC and enjoy it on VR devices through a WebXR application.

stilizer

stlizer/README.md

stlizer is a fast conservative video stabilizer.

cliqa

cliqa/README.md

cliqa provides low-vision image quality scores that are more consistent across different images.

It is useful for filtering low-quality images with a threshold value when creating image datasets.

Currently, the following two models are supported.

• JPEGQuality: Predicts JPEG Quality from image content
• GrainNoiseLeve: Predicts Noise Level related to photograph and PSNR degraded by that noise

CLI tools are also available to filter out low quality images using these results.

Install

Installer for Windows users

• nunif windows package
• nunif windows package (日本語)

For developers

Dependencies

• Python 3 (Works with Python 3.10 or later, developed with 3.10)
• PyTorch
• See requirements.txt

We usually support the latest version. If there are bugs or compatibility issues, we will specify the version.

• INSTALL-ubuntu
• INSTALL-windows
• INSTALL-macos

For Intel GPUs, additionally see section INSTALL-xpu.
For older NVIDIA GPUs, additionally see section INSTALL-cu126.

For container, packages, or special hardware builds, see extra_build.

About NUNIF_HOME

If the environment variable NUNIF_HOME is defined, downloaded pretrained models, configuration files, cache, temporary files, and lock files will be saved under NUNIF_HOME. This may be useful when packaging or in situations where the source directory does not have write permissions. The ~ character at the beginning of a path string is expanded to the home directory.

License Notes

Note that if you distribute binary builds, it is possible that it will be GPL.

This is due to PyAV(av) wheel package containing the GPL version of ffmpeg library. You can build PyAV with the LGPL version of ffmpeg library.

If you load this repository with torch.hub.load for waifu2x Python API etc, this problem does not exist because PyAV is not a dependent package.