# Module: tf.contrib.kfac.loss_functions

Loss functions to be used by LayerCollection.

## Classes

class CategoricalLogitsNegativeLogProbLoss: Neg log prob loss for a categorical distribution parameterized by logits.

class DistributionNegativeLogProbLoss: Base class for neg log prob losses that use the TF Distribution classes.

class LossFunction: Abstract base class for loss functions.

class MultiBernoulliNegativeLogProbLoss: Neg log prob loss for multiple Bernoulli distributions param'd by logits.

class NaturalParamsNegativeLogProbLoss: Base class for neg log prob losses whose inputs are 'natural' parameters.

class NegativeLogProbLoss: Abstract base class for loss functions that are negative log probs.

class NormalMeanNegativeLogProbLoss: Neg log prob loss for a normal distribution parameterized by a mean vector.

class NormalMeanVarianceNegativeLogProbLoss: Negative log prob loss for a normal distribution with mean and variance.

class OnehotCategoricalLogitsNegativeLogProbLoss: Neg log prob loss for a categorical distribution with onehot targets.

## Functions

insert_slice_in_zeros(...): Inserts slice into a larger tensor of zeros.