Module: tf.contrib.losses

View source on GitHub

Ops for building neural network losses.

See Contrib Losses.

Modules

metric_learning module: Ops for building neural network losses.

Functions

absolute_difference(...): Adds an Absolute Difference loss to the training procedure. (deprecated)

add_loss(...): Adds a externally defined loss to the collection of losses. (deprecated)

compute_weighted_loss(...): Computes the weighted loss. (deprecated)

cosine_distance(...): Adds a cosine-distance loss to the training procedure. (deprecated arguments) (deprecated)

get_losses(...): Gets the list of losses from the loss_collection. (deprecated)

get_regularization_losses(...): Gets the regularization losses. (deprecated)

get_total_loss(...): Returns a tensor whose value represents the total loss. (deprecated)

hinge_loss(...): Method that returns the loss tensor for hinge loss. (deprecated)

log_loss(...): Adds a Log Loss term to the training procedure. (deprecated)

mean_pairwise_squared_error(...): Adds a pairwise-errors-squared loss to the training procedure. (deprecated)

mean_squared_error(...): Adds a Sum-of-Squares loss to the training procedure. (deprecated)

sigmoid_cross_entropy(...): Creates a cross-entropy loss using tf.nn.sigmoid_cross_entropy_with_logits. (deprecated)

softmax_cross_entropy(...): Creates a cross-entropy loss using tf.nn.softmax_cross_entropy_with_logits. (deprecated)

sparse_softmax_cross_entropy(...): Cross-entropy loss using tf.nn.sparse_softmax_cross_entropy_with_logits. (deprecated)