ytm.guided.backends.act_loss.SigmoidBCE

class ytm.guided.backends.act_loss.SigmoidBCE(eps: float = 1e-7, pos_weights: list[float] | ndarray | None = None, neg_weights: list[float] | ndarray | None = None)

Bases: ActLoss

Sigmoid with binary cross-entropy per class, for binary and multi-label classification.

Parameters:

epsfloat, default=1e-7

Added inside the logs, must be > 0.

pos_weightsarray-like of float, default=None

Per-class weights of the positive term, one per class. None weights it by 1.

neg_weightsarray-like of float, default=None

Per-class weights of the negative term, one per class. None weights it by 1.