SGDGPU class

Applies the SGD optimizer to every element of its parameter tensors. SGD is a very simply descent method and my get stuck in local minima. This function does not require to be reallocated on each training step and can be reused indefinitely.

Inheritance

• Object
• OptimizerGPU
• SGDGPU

Constructors

SGDGPU(List<GPUTensor> parameters, double learningRate)

Initializes the SGD optimizer, which does not allocate any additional VRAM. Requires the list of trainable parameters and the learningRate.

Properties

hashCode → int

The hash code for this object.

no setter, inherited

learningRate ↔ double

getter/setter pair, inherited

parameters ↔ List<GPUTensor>

List of parameters the gradient is responsible for updating.

getter/setter pair, inherited

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

Methods

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

refreshStep(CommandBuffer tape) → void

Called once before every replay of an already-compiled tape. Optimizers with no per-step state (SGD) can leave this as a no-op; Adam overrides it to patch its step count.

inherited

step(CommandBuffer tape) → void

Appends the SGD weight update operations to the provided tape. This optimizer does not require to be updated and can be reused indefinitely via the same tape execution.

override

toString() → String

A string representation of this object.

inherited

zeroGrad(CommandBuffer tape) → void

Appends the zero-gradient operations to the tape to clear the gradients for the next pass.

override

Operators

operator ==(Object other) → bool

The equality operator.

inherited