VideoModelVariant

Enums · AiDotNet.Enums

Defines common model size variants for video processing models (SR, interpolation, flow, etc.).

For Beginners

Think of these like clothing sizes. “Tiny” is the fastest but least accurate, “Base” is the recommended starting point, and “Large”/“XLarge” are for when you need maximum quality and have powerful hardware.

How It Works

Video models typically come in multiple sizes trading off speed vs quality:

  • Smaller variants (Tiny, Small) run faster with lower memory, suitable for real-time applications
  • Larger variants (Large, XLarge) produce higher quality but require more compute
  • Base is the default recommended configuration balancing speed and quality

Fields

Field Summary
VideoModelVariant Base Base variant: default recommended configuration.
VideoModelVariant Large Large variant: increased capacity for higher quality.
VideoModelVariant Pro Pro variant: production-optimized variant with enhanced features.
VideoModelVariant Small Small variant: reduced size for faster inference (~20 FPS).
VideoModelVariant Tiny Tiny variant: minimal size for maximum speed (~30+ FPS).
VideoModelVariant XLarge Extra-large variant: maximum capacity.
int value__