std1.0.0-beta-7 SafeC docs

std :: ml::nn

Function std::tensor_huber_loss

#include <std/ml/nn.h> source ยท definition

&FloatTensor tensor_huber_loss(
    const &FloatTensor pred,
    const &FloatTensor target,
    float delta,
)

0.5 d^2 for |d| <= delta, else delta (|d| - delta/2). Gradient to pred only.