nll_loss_op

→TensorImplPtr

int nll_loss_op(const int & input, const int & target, const int & weight_or_null, int reduction, int ignore_index)

Public negative log-likelihood entry point.

inputTensorImplPtr

Log-probabilities of shape (N, C) or (N, C, d_1, ...).

targetTensorImplPtr

Integer class labels.

weight_or_nullTensorImplPtr

Optional per-class weight vector of length

C

; pass null when unused.

reductionint

0 (None), 1 (Mean), or 2 (Sum).

ignore_indexint

Class index excluded from the loss / gradient.

TensorImplPtr

Reduced or per-sample loss tensor.