Skip to main content

HuberLoss

HuberLoss(reduction="mean", delta=1.0)

Half the squared error below delta, a linear penalty above it, between input and target.

The module form of clika_runtime.nn.functional.huber_loss: construct it once with the settings, then call it like a function.

Args: reduction: how the elementwise losses reduce. delta: the switch point between the quadratic and the linear regime.

__init__​

__init__(self, reduction: '_Reduction' = 'mean', delta: 'float' = 1.0) -> 'None'

Initialize self. See help(type(self)) for accurate signature.

extra_repr​

extra_repr(self) -> 'str'

extra_repr() -> str

One line of per-class detail for :meth:__repr__; a layer prints its geometry here (in_features=64, out_features=256).

forward​

forward(self, input: 'Tensor', target: 'Tensor') -> 'Tensor'

forward(input, target) -> Tensor

Runs clika_runtime.nn.functional.huber_loss on input, target with the stored settings; returns a new tensor.