AI resizing for Vapoursynth using LIIF
Up- or downscaling to arbitrary resolutions and aspect ratios without ringing or blurriness. For example to go from 720x480 to 720x540, or to remove small black borders and stretch, or to downscale without resizing artifacts.
For large upscaling factors dedicated upscaling models are likely better.
- pytorch
pip install numpy
Put the entire vs_liif folder into your vapoursynth scripts folder.
Or install via pip: pip install git+https://github.com/pifroggi/vs_liif.git
import vs_liif
clip = vs_liif.resize(clip, width=720, height=540, src_left=0.0, src_top=0.0, device="cuda")
clip
Input clip must be in RGBS format.
width
Output width in pixel.
height
Output height in pixel.
src_left (optional)
Left shift in pixel. Allows for subpixel shift and negative shift.
src_top (optional)
Top shift in pixel. Allows for subpixel shift and negative shift.
device (optional)
Possible values are "cuda" to use with an Nvidia GPU, or "cpu". This will be extremely slow on CPU.
With large differences between input and output resolution, the liif model sometimes exhibits a small color shift. If you would like to undo this shift, try this: https://github.com/pifroggi/vs_colorfix
| Hardware | Resolution | Resize Factor | Average FPS |
|---|---|---|---|
| RTX 4090 | 720x480 | 0.25x (180x120) | ~30 fps |
| RTX 4090 | 720x480 | 0.5x (360x240) | ~20 fps |
| RTX 4090 | 720x480 | to 720x540 | ~10 fps |
| RTX 4090 | 720x480 | 1.5x (1080x720) | ~5 fps |
| RTX 4090 | 720x480 | 2x (1440x960) | ~3 fps |
| RTX 4090 | 1440x1080 | 0.25x (360x270) | ~5 fps |
| RTX 4090 | 1440x1080 | 0.5x (720x540) | ~3 fps |
| RTX 4090 | 1440x1080 | 1.5x (2160x1620) | ~0.8 fps |
| RTX 4090 | 1440x1080 | 2x (2880x2160) | ~0.5 fps |
Orignal code from "Learning Continuous Image Representation with Local Implicit Image Function" or LIIF.
Vapoursynth functions created with the help of ViktorThink.