This module provides different loss functions. A loss can be defined by a keyword (str) or a callable following the tensorflow.keras interface. If it is a keyword it has to be available in tensorflow.keras or in deephyper.losses. The loss functions availble in deephyper.losses are: * Negative Log Likelihood (compatible with Tensorflow Probability): tfp_negloglik or tfp_nll

deephyper.nas.losses.selectLoss(name: str)[source]

Return the loss defined by name.


name (str) – a string referenced in DeepHyper, one referenced in keras or an attribute name to import.


a string suppossing it is referenced in the keras framework or a callable taking (y_true, y_pred) as inputs and returning a tensor.

Return type

str or callable

deephyper.nas.losses.tfp_negloglik(y, rv_y)[source]

Negative log likelihood for Tensorflow probability.