public override compute(logits as NDArray<T>, labels as NDArray<Int>, gradientScale as Double = 1.0) as LossResult<T>
Computes mean loss and its mean-reduced output derivative.
| Parameter | Description |
|---|---|
logits | Matrix [samples, classes]. |
labels | Integer vector [samples]. |
gradientScale | Positive finite derivative multiplier; loss remains unscaled. |
Mean loss and gradient, without input mutation.
result = loss.compute(logits, labels)