Crate amari_dual 

Dual number automatic differentiation for efficient gradient computation

Dual numbers extend real numbers with an infinitesimal unit ε where ε² = 0. This allows for exact computation of derivatives without numerical approximation or computational graphs, making it ideal for forward-mode automatic differentiation.

Re-exports

pub use error::DualError;

pub use error::DualResult;

pub use multivector::DualMultivector;

pub use multivector::MultiDualMultivector;

pub use types::DualNumber;

pub use types::MultiDualNumber;

pub use types::StandardDual;

pub use types::StandardMultiDual;

pub use types::ExtendedDual;

pub use types::ExtendedMultiDual;

Modules

error

Error types for dual number operations

functions

Common functions with automatic differentiation support

multivector

Dual number multivectors for automatic differentiation in geometric algebra

types

Core dual number types for automatic differentiation

Traits

PrecisionFloat

Unified trait for floating-point arithmetic in scientific computing

Type Aliases

ExtendedFloat

HighPrecisionFloat

StandardFloat

Type alias for standard precision calculations