TorchLean API

Docs HomeInstallationGuideExamplesGraphsTools

FloatLib.Floats.Formats.FixedPoint.Bounded.Configured.Conversion.Instances

Bounded fixed-point conversion instances #

Bounded fixed-point destinations support exact-rational quantization under an explicit overflow policy. The context-free instance chooses checked conversion and therefore rejects overflow.

source
@[instance_reducible]
instance FloatLib.Floats.ExecFloat.BoundedFixedPoint.Conversion.quantizer {radix : Numerics.Radix} {fractionalDigits width : ℕ} :
Quantizer (BoundedFixedPoint radix fractionalDigits width) ℚ

Bounded fixed-point conversion requires an explicit overflow policy.

source
@[instance_reducible]
instance FloatLib.Floats.ExecFloat.BoundedFixedPoint.Conversion.defaultQuantizer {radix : Numerics.Radix} {fractionalDigits width : ℕ} :
DefaultQuantizer (BoundedFixedPoint radix fractionalDigits width) ℚ

Context-free bounded conversion is checked and rejects overflow.