cross_entropy_op

→TensorImplPtr

int cross_entropy_op(const int & input, const int & target, const int & weight_or_null, int reduction, double eps, int ignore_index)

Public cross-entropy (fused log-softmax + NLL) entry point.

inputTensorImplPtr

Logits of shape (N, C) or (N, C, d_1, ...).

targetTensorImplPtr

Integer class labels of shape (N,) or (N, d_1, ...).

weight_or_nullTensorImplPtr

Optional per-class weight vector of length

C

; pass null when unused.

reductionint

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

epsdouble

Numerical clamp passed through to the backend.

ignore_indexint

Class index excluded from the loss and gradient.

TensorImplPtr

Scalar (Mean / Sum) or per-sample (None) loss tensor.