xlnstorch.operators.lns_cross_entropy

xlnstorch.operators.lns_cross_entropy(x, target, base, weight=None, size_average=None, ignore_index=-100, reduce=None, reduction='mean', label_smoothing=0.0)

See docs for torch.nn.functional.cross_entropy() for parameter/return details. Typically, torch.Tensor arguments are the equivalent internal representations of LNStensors.

Parameters:

• x (Tensor)

• target (Tensor)

• base (Tensor)

• weight (Tensor)

• size_average (bool)

• ignore_index (int)

• reduce (bool)

• reduction (str)

• label_smoothing (float)

Return type:

Tensor