Log-cosh loss is a loss function that is used to solve regression problems. Log-cosh is calculated as the average logarithm of the hyperbolic cosine of the differences between the predicted and actual values.
The formula to calculate the log-cosh:
n- the number of data points.
y- the actual value of the data point. Also known as true value.
ŷ- the predicted value of the data point. This value is returned by model.
Let's say we have the following sets of numbers:
|actual values of ||4||-1.5||5||2|
|predicted values of ||3.5||1||5||3|
Here is example how log-cosh can be calculated using these numbers:
TensorFlow 2 allows to calculate the log-cosh. It can be done by using
from tensorflow import keras yActual = [4, -1.5, 5, 2] yPredicted = [3.5, 1, 5, 3] logcoshObject = keras.losses.LogCosh() logcoshTensor = logcoshObject(yActual, yPredicted) logcosh = logcoshTensor.numpy() print(logcosh)
Log-cosh also can be calculated by using
logcoshTensor = keras.losses.log_cosh(yActual, yPredicted) logcosh = logcoshTensor.numpy()