TorchLean API

FloatLib.Floats.Formats.IEEE754.Native.Info

Inspection of Lean's native runtime floats #

#float_info reports native storage, arithmetic, and conversions for Lean's Float32 and binary64 Float. It lists the Lean 4.34 logical-model bridges with their input conditions and the certified software refinements for their configured counterparts. Execution of native arithmetic, including explicit calls through Configured.NativeFPU.Unchecked, also depends on Lean's compiler, runtime, and host hardware.

def FloatLib.Floats.Formats.IEEE754.Native.FloatInfo.profile (bits exponentBits fractionBits : ) (name storage : String) :

Storage, operations, and conversion theorems for a Lean native float type.

Instances For