pub struct RCGConfig {
pub max_iters: usize,
pub grad_tol: Real,
pub init_step: Real,
pub armijo_c: Real,
pub armijo_beta: Real,
pub max_ls_iters: usize,
pub variant: CgVariant,
pub restart_every: usize,
}Expand description
Configuration for Riemannian Conjugate Gradient.
Fields§
§max_iters: usizeMaximum number of iterations.
grad_tol: RealStop when ||grad f(x)|| < grad_tol.
init_step: RealInitial step size for Armijo backtracking.
armijo_c: RealArmijo sufficient decrease constant.
armijo_beta: RealBacktracking factor (< 1).
max_ls_iters: usizeMaximum Armijo backtracking steps per iteration.
variant: CgVariantFletcher-Reeves or Polak-Ribière+.
restart_every: usizeRestart to steepest descent every N iterations (0 = never force restart).
Automatic restart still occurs when the conjugate direction is not a descent direction.
Trait Implementations§
Auto Trait Implementations§
impl Freeze for RCGConfig
impl RefUnwindSafe for RCGConfig
impl Send for RCGConfig
impl Sync for RCGConfig
impl Unpin for RCGConfig
impl UnsafeUnpin for RCGConfig
impl UnwindSafe for RCGConfig
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more