Loss Forward Layer

The loss function of a neural network is the sum of loss functions for each data sample. The loss function measures and penalizes the difference between the output of the neural network and ground truth. Because the loss layer evaluates the quality of the model being trained, it must be the last layer in the neural network topology.

The loss layer takes the ground truth T and output values X of the preceding layer as input and computes the value y of the cost function E:



For more complete information about compiler optimizations, see our Optimization Notice.
Select sticky button color: 
Orange (only for download buttons)