waifu2x-ncnn-vulkan nihui/waifu2x-ncnn-vulkan

Portable command-line waifu2x upscaler and denoiser that runs on Intel, AMD, Nvidia and Apple Silicon GPUs via Vulkan.

Creator kit + Compare GitHub ↗ Install guide (official) ↗

waifu2x-ncnn-vulkan at a glance

Checked 16 Oct 2026

A fast ncnn/Vulkan port of waifu2x, the classic upscaler for anime-style art, bundled with CUNet, anime-art and photo models. It upscales 2x to 32x, removes JPEG noise at four levels, and accepts a whole directory as input. Portable builds for Windows, Linux and macOS need no CUDA; a CPU mode exists. Code and bundled waifu2x models are MIT. It is command-line only (GUIs such as Waifu2x-Extension-GUI wrap it), and on real photos it is weaker than GAN models like Real-ESRGAN. Last release Sep 2025.

Cite this pageRepoQuarry. “waifu2x-ncnn-vulkan.” Checked 16 Oct 2026. https://repoquarry.dev/projects/waifu2x-ncnn-vulkan

How much of Topaz Gigapixel does it replace?

4 of 6 jobs. Strongest on anime and illustrations; JPEG denoise and folder batches from the command line; no GUI or face recovery.

Photo upscaling up to 16x Partial replacement

Face recovery on blurry or low-res faces Not covered

Recover and redefine low-quality or compressed images Partial replacement

Upscaling art Partial replacement

Batch upscaling of many images Partial replacement

Photoshop and Lightroom plugin Not covered

For face recovery on blurry or low-res faces, Photoshop and Lightroom plugin, see other Topaz Gigapixel alternatives. You may also need to build an interface, an API server and hosting yourself.

Facts, with sources

Every answer links to where we found it. We read the documentation; we did not install it.

Report a correction

QuestionAnswerSourceChecked

Runs on my machine? Yes, portable binary This package includes all the binaries and models required. It is portable, so no CUDA or Caffe runtime environment is needed :) github.com · waifu2x-ncnn-vulkan#download ↗ today

Code licence MIT github.com · LICENSE ↗ today

Model weights licence MIT; bundled models come from the original waifu2x project, which is MIT github.com · LICENSE ↗ today

Commercial use Yes, MIT code and models github.com · LICENSE ↗ today

Paid services needed? None required github.com · waifu2x-ncnn-vulkan#readme ↗ today

Hardware Vulkan GPU (Intel, AMD, Nvidia, Apple Silicon); CPU mode with -g -1; smaller tiles cut GPU memory use -g gpu-id gpu device to use (-1=cpu, default=auto) can be 0,1,2 for multi-gpu github.com · waifu2x-ncnn-vulkan#full-usages ↗ today

Operating systems Windows, Linux, macOS Download Windows/Linux/macOS Executable for Intel/AMD/NVIDIA/Apple-Silicon GPU github.com · waifu2x-ncnn-vulkan#download ↗ today

Setup effort Easy, unzip and run from the command line waifu2x-ncnn-vulkan.exe -i input.jpg -o output.png -n 2 -s 2 github.com · waifu2x-ncnn-vulkan#example-command ↗ today