Losses
A loss function (or objective function), specified when you compile the model, is the function that the model intends to optimize in the process of training.
See here for available loss objects.
For the sake of convenience, you can also use the corresponding string representation of a loss.
Available Losses
- mean_squared_error or mse
- mean_absolute_error or mae
- categorical_crossentropy
- sparse_categorical_crossentropy
- binary_crossentropy
- mean_absolute_percentage_error or mape
- mean_squared_logarithmic_error or msle
- kullback_leibler_divergence or kld
- hinge
- squared_hinge
- poisson
- cosine_proximity