losses (pmrf.losses)

Loss models for optimizer fitting or generalized Bayesian inference.

These classes wrap pure mathematical loss functions into a pmrf.Loss. All losses take the true and predict arrays as inputs, and return the loss value when called.

Classes

AbstractLoss()	Abstract base class for frequentist loss functions.
LogMSELoss([multioutput])	Log of Mean Squared Error (RMSE) metric.
RMSELoss([multioutput])	Root Mean Squared Error (RMSE) metric.
MAPELoss([multioutput])	Mean Absolute Percentage Error (MAPE) metric.
HuberLoss([delta, multioutput])	Huber loss metric.
HingeLoss([operator, weight, mask, ...])	Applies a one-sided constraint (hinge) before evaluating a base metric.