video2x

Project Overview

Video2X is a machine learning-based video super resolution and frame interpolation framework. Version 6.0.0 is a complete rewrite in C/C++, supporting Windows and Linux with faster performance, higher efficiency, and significantly improved output quality.

Key Features

• Super resolution: Supports Anime4K v4, Real-ESRGAN, Real-CUGAN
• Frame interpolation: Supports RIFE (all models)
• Two modes: Filtering (upscaling) and frame interpolation
• Cross-platform: Supports both Windows and Linux
• GUI: Easy-to-use GUI and installer for Windows
• Zero extra disk space: No additional disk space required during processing

Target Audience

• Creators and enthusiasts who need to enhance video quality
• Users who want to upscale low-resolution videos
• Users who need frame interpolation

Quick Start

• Windows: Download the installer video2x-qt6-windows-amd64-installer.exe
• Linux: Arch Linux users can install from AUR (video2x, video2x-git, video2x-qt6)
• Other distributions can use the AppImage

Core Advantages

• Machine learning-based super resolution and frame interpolation
• Supports Anime4K v4, Real-ESRGAN, Real-CUGAN, and RIFE models
• C/C++ rewrite for faster and more efficient performance
• Cross-platform support for Windows and Linux

System Requirements

• CPU: Requires AVX2 support (Intel Haswell Q2 2013 or newer, AMD Excavator Q2 2015 or newer)
• GPU: Requires Vulkan support (NVIDIA Kepler GTX 600 series or newer, AMD GCN 1.0 or newer, Intel HD Graphics 4000 or newer)