Adam class

Inheritance

Constructors

Adam(List<Tensor> parameters, {double lr = 1e-3, double beta1 = 0.9, double beta2 = 0.999, double eps = 1e-8, double weightDecay = 0.0})

Properties

beta1 → double
final
beta2 → double
final
eps → double
final
hashCode → int
The hash code for this object.
no setterinherited
lr ↔ double
Current learning rate. Mutable so that LRSchedulers can adjust it across training steps.
getter/setter pairoverride
parameters → List<Tensor>
The parameters this optimizer updates. Order is stable.
finalinherited
runtimeType → Type
A representation of the runtime type of the object.
no setterinherited
weightDecay → double
final

Methods

noSuchMethod(Invocation invocation) → dynamic
Invoked when a nonexistent method or property is accessed.
inherited
step() → void
Apply one update step using each parameter's current .grad. Parameters with grad == null are skipped.
override
toString() → String
A string representation of this object.
inherited
zeroGrad() → void
Zero the gradient buffer of every registered parameter.
inherited

Operators

operator ==(Object other) → bool
The equality operator.
inherited