xma.optimizers.sgd

class SGD(params: Iterable[Tensor] | Iterable[dict[str, Any]] | Iterable[tuple[str, Tensor]], lr: float | Tensor = 0.001, momentum: float = 0, dampening: float = 0, weight_decay: float | Tensor = 0, nesterov: bool = False, *, maximize: bool = False, foreach: bool | None = None, differentiable: bool = False, fused: bool | None = None)

Bases: SGD

step(closure: Callable | None = None, *, kernel_backend: KernelBackend | None = None) → None

Perform a single optimization step.

Parameters:

closure (Callable, optional) – A closure that reevaluates the model and returns the loss.

sgd(params: list[Tensor], grads: list[Tensor], momentum_buffer_list: list[Tensor], lr: float, weight_decay: float, momentum: float, dampening: float, nesterov: bool, maximize: bool, foreach: bool, *, kernel_backend: KernelBackend | None = None) → None