Home / Tools / XVFI

XVFI

AI frame interpolation for ultra-smooth slow motion

Visit XVFI
XVFI preview
github.com Free

Upvotes

0

Pricing

Free

Category

Video Enhancers

In the database since

Oct 2026

Overview

About XVFI

XVFI is a free research project for extreme video frame interpolation with artificial intelligence. It is the official code for an ICCV 2021 oral paper. The method creates in-between frames, turning ordinary footage into ultra-smooth high-frame-rate video. It targets researchers and developers comfortable with Python and PyTorch. Training code, test code, trained weights, and datasets are all provided. Frame interpolation means the AI draws new frames between the ones your camera captured. Check the official GitHub page for current details and setup guides.

Key Features

1. Extreme Frame Interpolation

Generates many new frames between each pair of originals. It can push video toward 1,000 frames per second for dramatic slow motion.

2. 4K Support

Handles ultra-high-definition 4K footage, not just small clips. The X-TEST set includes extended 4K sequences for testing.

3. Complete Code Release

Includes training code, testing code, and the trained weights. The same datasets from the paper are shared for reproduction.

4. Paper and Demos

Links to the paper, a demo video, and the oral presentation are in the repo. Visual results help you judge quality before running code.

5. GUI Option

The authors link to Flowframes, a graphical app built on similar ideas. It offers frame interpolation without touching code.

Frequently Asked Questions (FAQ)

What is video frame interpolation?

Creating new frames between existing ones to raise the frame rate. XVFI uses AI to draw these middle frames so motion looks smoother.

What makes XVFI extreme?

It targets very large frame-rate jumps, tested up to 1,000 fps on 4K footage. Most interpolators only double or quadruple the frame rate.

Do I need a GPU?

Yes, a GPU is recommended for practical use. Frame interpolation on 4K video is heavy work for a CPU alone.

Can I try XVFI without coding?

The authors link to Flowframes, a graphical interface for frame interpolation. The GitHub repo itself needs Python and PyTorch.

Where can I see results first?

The repo links a YouTube demo and the ICCV oral presentation video, so you can judge the output quality before installing anything.

Conclusion

XVFI is a powerful research tool for pushing video to extreme frame rates with AI. It suits developers and researchers working on slow motion and video enhancement. Full code, weights, and datasets are shared openly. Check the official GitHub page for the paper, demos, and installation steps. Its ICCV 2021 oral selection marks it as a notable advance.

Newsletter

Tools like this change fast

New Video Enhancers tools launch every week — the newsletter keeps you ahead of what's worth trying.

Keep exploring

Similar AI tools

Ranked by the job they do, their capabilities, and how closely they fit this workflow.