Skip to content
waifu2x by w2xc for VapourSynth
Branch: master
Clone or download
Latest commit 4128c53 May 19, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
Waifu2x-w2xc Update models May 19, 2018
.gitattributes Initial commit Jul 20, 2015
.gitignore Initial commit Jul 20, 2015
LICENSE Replace custom configure script with Autotools May 19, 2018 Update May 19, 2018 Set executable bit on May 19, 2018 Replace custom configure script with Autotools May 19, 2018


waifu2x filter for VapourSynth, based on the w2xc library.


The models folder must be located in the same folder as Waifu2x-w2xc.dll.

The filter will generate .bin files at the same location of the model files for the first time it runs. Make sure that your executable has write permission to the folder of the model files. The filter still can run even without the .bin files generated, but the performance will degrade. When you update the model files in the future, remember to delete the .bin files as well.


w2xc.Waifu2x(clip clip[, int noise=1, int scale=2, int block=512, bint photo=False, int gpu=1, int processor=-1, bint list_proc=False, bint log=False])
  • clip: Clip to process. Only planar format with float sample type of 32 bit depth is supported.

  • noise: Noise reduction level.

    • 0 = none
    • 1 = medium
    • 2 = high
    • 3 = highest
  • scale: Upscaling factor. Must be a power of 2. Set to 1 for no upscaling.

  • block: The block size for dividing the image during processing. Smaller value results in lower VRAM usage, while larger value may give faster speed. The optimal value may vary according to different graphics card and image size.

  • photo: When set to false, it uses the anime-style model. When set to true, the photo model will be used instead. There are two sets of anime-style models, one is trained under RGB, the other is trained under Y (luma only). The RGB model will be used when the color family of the input is RGB, and the Y model will be used for all the other color family. Note that the photo model is only available for RGB, so this parameter has no effect for non-RGB input.

  • gpu: Controls the environment to use.

    • 0 = disable GPU
    • 1 = auto detect. It will run on the first available environment in the following order:
      • CUDA
      • AMD OpenCL
      • FMA
      • AVX
      • Intel OpenCL
      • SSE3
      • OpenCV filter2D
    • 2 = force to use OpenCL
  • processor: Sets target processor. Any non-negative value supersedes the choice from gpu. Use list_proc to get the index of the available processors.

  • list_proc: Whether the processors list is drawn on the frame.

  • log: Whether the internal processing stats are written to stdout.


Requires w2xc. The script doesn't copy the models folder, hence you have to do it yourself.

You can’t perform that action at this time.