Enum GradientStrategy 

pub enum GradientStrategy {
    None,
    Full,
    Checkpointed {
        checkpoint_interval: usize,
    },
    Accumulated {
        accumulation_steps: usize,
    },
}

Gradient computation strategy

Variants

None

No gradient computation

Full

Full gradient computation (standard backprop)

Checkpointed

Checkpointing (recompute forward pass to save memory)

Fields

checkpoint_interval: usize

Checkpoint every N nodes

Accumulated

Gradient accumulation across multiple steps

Fields

accumulation_steps: usize

Number of steps to accumulate before updating

Trait Implementations

impl Clone for GradientStrategy

fn clone(&self) -> GradientStrategy

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for GradientStrategy

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for GradientStrategy

fn eq(&self, other: &GradientStrategy) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for GradientStrategy

impl Eq for GradientStrategy

impl StructuralPartialEq for GradientStrategy