LogCoshLoss<T>

Models & Types · AiDotNet.LossFunctions

Implements the Log-Cosh loss function, a smooth approximation of Mean Absolute Error.

Inheritance: LossFunctionBase<T> → LogCoshLoss<T>

For Beginners

Log-Cosh loss is a smooth approximation of the Mean Absolute Error. It calculates the logarithm of the hyperbolic cosine of the difference between predictions and actual values.

This loss function has several desirable properties:

  • It’s smooth everywhere (unlike Huber loss which has a point where its derivative is not continuous)
  • It’s less affected by outliers than Mean Squared Error
  • It behaves like Mean Squared Error for small errors and Mean Absolute Error for large errors
  • Its derivatives are well-defined and bounded, which helps prevent gradient explosions during training

Log-Cosh loss is particularly useful for regression problems where:

  • You want the smoothness of MSE but with better robustness to outliers
  • You need stable gradients for model training
  • You want a compromise between MSE and MAE

Example

using AiDotNet.LossFunctions;
using AiDotNet.Tensors.LinearAlgebra;

var loss = new LogCoshLoss<float>();
var predicted = new Vector<float>(new[] { 0.9f, 0.2f, 0.7f });
var actual = new Vector<float>(new[] { 1.0f, 0.0f, 1.0f });

float value = loss.CalculateLoss(predicted, actual);
Console.WriteLine($"LogCoshLoss = {value:F4}");

Constructors

Constructor Summary
LogCoshLoss()

Methods

Method Summary
T CalculateLoss(Vector<T> predicted, Vector<T> actual) Calculates the Log-Cosh loss between predicted and actual values.
Tensor<T> ComputeTapeLoss(Tensor<T> predicted, Tensor<T> target)
ValueTuple<T, Tensor<T>> CalculateLossAndGradientGpu(Tensor<T> predicted, Tensor<T> actual) Calculates both Log-Cosh loss and gradient on GPU in a single efficient pass.