FluidFrames.RIFE | video AI frame generation app

What is FluidFrames.RIFE?

FluidFrames.RIFE is a Windows app powered by RIFE AI to create frame-generated and slowmotion videos.

Other AI projects.🤓

• https://github.com/Djdefrag/QualityScaler / QualityScaler - image/video AI upscaler app
• https://github.com/Djdefrag/RealScaler / RealScaler - image/video AI upscaler app (Real-ESRGAN)

Credits.

• RIFE - https://github.com/megvii-research/ECCV2022-RIFE
• PraticalRIFE - https://github.com/hzwer/Practical-RIFE

How is made. 🛠

FluidFrames.RIFE is completely written in Python, from backend to frontend. External packages are:

• AI -> torch / onnxruntime-directml
• GUI -> customtkinter
• Image/video -> OpenCV / moviepy
• Packaging -> Nuitka

Requirements. 🤓

• Windows 11 / Windows 10
• RAM >= 8Gb
• Any Directx12 compatible GPU with >= 2GB VRAM

Features.

• Elegant and easy to use GUI
• Resize video before interpolation
• Multiple GPUs support
• Compatible video - mp4, wemb, gif, mkv, flv, avi, mov, qt
• Video frame-generation STOP&RESUME
• PRIVACY FOCUSED - no internet connection required / everything is on your PC
• Video frames generation x2 / x4 / x8
  • 30fps => x2 => 60fps
  • 30fps => x4 => 120fps
  • 30fps => x8 => 240fps
• Video slowmotion x2 /x4
  • 30fps => x2_slowmotion => 30fps - 2 times slower
  • 30fps => x4_slowmotion => 30fps - 4 times slower
  • 30fps => x8_slowmotion => 30fps - 8 times slower

Next steps. 🤫

• 1.X versions
  • Switch to Pytorch-directml to support all Directx12 compatible gpu (AMD, Intel, Nvidia)
  • New GUI with Windows 11 style
  • Include audio for processed video
  • Optimizing video frame resize and extraction speed
  • Multi GPU support (for pc with double GPU, integrated + dedicated)
  • Python 3.10 (expecting ~10% more performance)
  • Slowmotion function
• 2.X versions
  • New, completely redesigned graphical interface based on @customtkinter
  • Fluidify multiple videos at once
  • Save AI generated frames as files
  • Support RIFE AI model updates
  • Support for RIFE_Lite AI model (a faster and lighter version of RIFE)
• 3.x versions (now under development)
  • New AI engine powered by onnxruntime-directml (https://github.com/microsoft/onnxruntime)
  • Python 3.11 (~10% performance improvements)
  • Display frame-generated videos info in the GUI
  • FFMPEG 7 (latest release)
  • Saving user settings (AI model, GPU, CPU etc.)
  • Video frame-generation STOP&RESUME
  • Python 3.12

Some Examples.

Videos

1. Original / x4 / x2-slomotion

giphy_RIFEx4_100.mp4

giphy_RIFEx2_slowmo_100.mp4

3. Original / x4 / x4-slomotion

deadpool.mp4

deadpool_RIFEx4_100.mp4

deadpool_RIFEx4_slowmo_100.mp4

3. Original / x2

original.mp4

gg.ss.RIFEHDv3.mp4

4. Original / x2

209639439-94c8774d-354e-4d56-9123-e1aa4af95e08_RIFE_HDv3.mp4

5. Original / x2 / x2-slomotion

gohan.mp4

gohan_RIFE_HDv3x2.mp4

gohan_RIFE_HDv3x2-slowmotion.mp4