FoundationModelSize

Enums · AiDotNet.Enums

Defines the size variants available for foundation models.

For Beginners

Larger models generally produce better predictions but require more memory and computation time. Choose a size based on your resources:

  • Tiny/Mini: Fast experiments, limited hardware, edge deployment
  • Small/Base: Good balance of quality and speed for most use cases
  • Large/XLarge: Best accuracy when compute resources are available

How It Works

Foundation models typically come in multiple sizes, trading off between accuracy and computational cost. This enum replaces error-prone string-based size selection (e.g., "base", "large") with compile-time type safety.

Fields

Field Summary
FoundationModelSize Base Base variant (~50-200M parameters).
FoundationModelSize Large Large variant (~200-700M parameters).
FoundationModelSize Mini Mini variant (~5-20M parameters).
FoundationModelSize Small Small variant (~14-50M parameters).
FoundationModelSize Tiny Tiny variant with minimal parameters (~1-5M).
FoundationModelSize XLarge Extra-large variant (700M+ parameters).
int value__