LogCosh

public class LogCosh

Computes Computes the logarithm of the hyperbolic cosine of the prediction error.

logcosh = log((exp(x) + exp(-x))/2), where x is the error predictions - labels.

Standalone usage: 

    Operand<TFloat32> labels =
        tf.constant(new float[][] { {0.f, 1.f}, {0.f, 0.f} });
    Operand<TFloat32> predictions =
        tf.constant(new float[][] { {1.f, 1.f}, {0.f, 0.f} });
    LogCosh logcosh = new LogCosh(tf);
    Operand<TFloat32> result = logcosh.call(labels, predictions);
    // produces 0.108

Calling with sample weight: 

    Operand<TFloat32> sampleWeight = tf.constant(new float[] {0.8f, 0.2f});
    Operand<TFloat32> result = logcosh.call(labels, predictions, sampleWeight);
    // produces 0.087f

Using SUM reduction type: 

    LogCosh logcosh = new LogCosh(tf, Reduction.SUM);
    Operand<TFloat32> result = logcosh.call(labels, predictions);
    // produces 0.217f

Using NONE reduction type: 

    LogCosh logcosh = new LogCosh(tf, Reduction.NONE);
    Operand<TFloat32> result = logcosh.call(labels, predictions);
    // produces [0.217f, 0f]

Inherited Fields

Public Constructors

LogCosh(Ops tf)
Creates a LogCosh Loss using getSimpleName() as the loss name and a Loss Reduction of REDUCTION_DEFAULT
LogCosh(Ops tf, String name)
Creates a LogCosh Loss using a Loss Reduction of REDUCTION_DEFAULT
LogCosh(Ops tf, Reduction reduction)
Creates a LogCosh Loss using getSimpleName() as the loss name
LogCosh(Ops tf, String name, Reduction reduction)
Creates a LogCosh Loss

Public Methods

<T extends TNumber> Operand<T>
call(Operand<? extends TNumber> labels, Operand<T> predictions, Operand<T> sampleWeights)
Generates an Operand that calculates the loss.

Inherited Methods

Public Constructors

public LogCosh (Ops tf)

Creates a LogCosh Loss using getSimpleName() as the loss name and a Loss Reduction of REDUCTION_DEFAULT

Parameters
tf the TensorFlow Ops

public LogCosh (Ops tf, String name)

Creates a LogCosh Loss using a Loss Reduction of REDUCTION_DEFAULT

Parameters
tf the TensorFlow Ops
name the name of the loss, if null then getSimpleName() is used.

public LogCosh (Ops tf, Reduction reduction)

Creates a LogCosh Loss using getSimpleName() as the loss name

Parameters
tf the TensorFlow Ops
reduction Type of Reduction to apply to the loss.

public LogCosh (Ops tf, String name, Reduction reduction)

Creates a LogCosh Loss

Parameters
tf the TensorFlow Ops
name the name of the loss, if null then getSimpleName() is used.
reduction Type of Reduction to apply to the loss.

Public Methods

public Operand<T> call (Operand<? extends TNumber> labels, Operand<T> predictions, Operand<T> sampleWeights)

Generates an Operand that calculates the loss.

Parameters
labels the truth values or labels
predictions the predictions
sampleWeights Optional sampleWeights acts as a coefficient for the loss. If a scalar is provided, then the loss is simply scaled by the given value. If SampleWeights is a tensor of size [batch_size], then the total loss for each sample of the batch is rescaled by the corresponding element in the SampleWeights vector. If the shape of SampleWeights is [batch_size, d0, .. dN-1] (or can be broadcast to this shape), then each loss element of predictions is scaled by the corresponding value of SampleWeights. (Note on dN-1: all loss functions reduce by 1 dimension, usually axis=-1.)
Returns
  • the loss