REAL Video Enhancer 2.2.0 Pre-Release

RVE 2.2.0 pre-release

• NOTE: Pre-releases are unstable, please use the stable build if you experience issues.
  New features will be added to this release over time, the current changelog is not final.

Added

• Hardware Encoding
• Auto border cropping
• GPU ID selection
• Default video container setting
• Batch input support

Changed

• Adjusted dynamic scale.
• Transcodes audio by default, to fix potential audio copying issues.
• Moved pausing to shared memory.

Fixed

• Win10 tooltips and pop ups.

REAL Video Enhancer 2.1.5

Discord: https://discord.gg/YHM7nFheyM

Added

• Stopping render, instead of having to kill the entire app.
• More tiling options.
• Better checks on imported models.
• Subtitle passthrough.
• Changelog view in home menu.
• Upscale and Interpolate at the same time.
• Sudo shuffle span for pytorch backend
• GIMM-VFI
• GMFSS Pro, which helps fix text warping.
• SloMo mode
• Ensemble for Pytorch/TensorRT interpolation.
• Dynamic Optical Flow for Pytorch interpolation.

Changed

• Make RVE smaller by switching to pyside6-essentials. (thanks zeptofine!)
• Make GUI more compact.
• Bump torch to 2.6.0-dev20241214.
• Bump CUDA to 12.6
• Remove System CUDA install requirement for GMFSS

Fixed

• ROCm upscaling not working.

Known Issues

• Windows detects RVE as a trojan. This is a false positive, if you do not trust the build, instructions on how to build are in the readme.
• Linux xorg not launching due to libxcb-cursor0 not being installed, please install this via your package manager.
• There are a lot of changes in this release, there may be bugs I didn't catch. Please create issues if you find such bugs.

REAL Video Enhancer 2.1.0

Discord: https://discord.gg/YHM7nFheyM

Added

• Custom Upscale Model Support (TensorRT/Pytorch/NCNN)
• GMFSS
• RIFE 4.25 (TensorRT/Pytorch/NCNN)
• PySceneDetect for better scene change detections
• More NCNN models

Removed

• MacOS Support fully, not coming back due to changes made by Apple.

Changed

• Simplified TensorRT Engine Building
• Increased RIFE TensorRT speed
• Better preview that pads instead of stretches
• Updated PyTorch to 2.6.0.dev20241023
• Updated TensorRT to 10.6
• Naming scheme of upscaling models, should be easier to understand

REAL Video Enhancer 2.0.5

Hotfix release

• Adds
  • Option to uninstall a specific backend (Experimental)
• Fixes
  • SPAN incorrect output and broken TRT
  • Rife 4.6 TRT

REAL Video Enhancer 2.0

Added

• 🪟 Windows support: Now fully compatible with Windows OS.
• 🌐 Whole new frontend and backend.
• 📂 Custom output file location.
• 🆕 Many new RIFE models.
• 🔍 More Upscaling models.
• 🔧 Selectable backends on install.
• ⏸️ Pausing render to free GPU usage on long inference.
• 🐛 Various bug fixes.

Upgraded

• 🐍 Base backend now uses portable Python, rather than PyInstaller.
• 🚀 Updated TensorRT to 10.3.
• 🔄 Updated PyTorch to 2.5.0.
• ⚡ New RIFE implementation, faster and more accurate!

Removed

• 🍏 MacOS support (Will soon be back!)
• 🗃️ Custom model importing.
• 📹 Interpolating from YouTube video.

Hotpatches:

1. Fixed backend not downloading
2. Fixed crash if DiscordRPC was not enabled

REAL Video Enhancer 2.0 Alpha 4

Adds:

• Pausing
• Custom model imports (broken for now)
• More settings
• Bugs

Fixes:

• TensorRT RIFE (Thanks Holywu)
• a lot

REAL Video Enhancer 2.0 Alpha 3

Fixes:

• Rife interpolation at a high multiplier now calculates timestep correctly
• Output res gets recalculated when switching upscale model

REAL Video Enhancer 2.0 Alpha 2

Fixes

• crash when discord isn't running on Windows
• race condition when upscaling using tensorrt

REAL Video Enhancer 2.0 Alpha 1

Additions:

• Windows Support
• Whole new frontend and backend.
• Custom output file location
• Many new RIFE models
• more Upscaling models
• Selectable backends on install
• Bugs

Changes:

• Base backend on portable python, rather than pyinstaller.
• Update TensorRT to 10.3
• Update PyTorch to 2.5.0-dev

REAL Video Enhancer 1.2.0

Discord: https://discord.gg/YHM7nFheyM

Additions:

• TensorRT support (8.6.0)
  • Rife TensorRT (1080p max recommended input)
  • Upscaling TensorRT (Custom models + Esrgan, some custom models might not work!)
• Adds option to disable frame extraction in rife-ncnn
• SPAN-ncnn
• Bugs (confirmed)

Changes:

• Swapped CX_Freeze to PyInstaller for CUDA build, please report issues that come up.
• Performance improvements across the board.
• Auto detect half precision mode, so supported GPUs will have it enabled by default. (Broken)
• Disable half precision button if your gpu does not support it. (Broken)

Fixes:

• Fix GMFSS (Broken)

Known Issues:

• CUPY does not recognize cuda install, so GMFSS is still broken.
• Auto detect half precision does not work.
• Inference with certain file names can stop scene change detection from working.
• Some distros have issues with TensorRT inference, please revert back to CUDA if it crashes.

Credits:

• https://github.com/NevermindNilas/TheAnimeScripter/ (For TensorRT implementation and some optimizations)
• https://github.com/adegerard (For many TensorRT optimizaitons)

Download Guide:

• For NVIDIA users, download the CUDA tar.gz file. Extract it, Note: Make sure all the files are together, THIS IS NOT AN APPIMAGE.
  • Support for Ubuntu 22.04+ on CUDA edition because of packaging issues, NCNN version still supports 20.04+
• For MacOS users, refer to here
• For AMD and Intel GPU users, download the NCNN AppImage. Nothing has changed, fuse2 is required to launch the AppImage.