Terminal layer of a CeNiN network. It does not compute activations; it exposes the class labels together with the probabilities produced by the preceding softmax layer.
Output
01 Syntax
02 Methods
| Name | Overloads | Summary |
|---|---|---|
| .ctor | 1 | Creates the output layer. |
| getDecision | 1 | Sorts the class probabilities in descending order and returns the most likely class label. |
| layerFeedNext | 1 | Always throws because the output layer is the end of the network and cannot feed another layer. |
| feedNext | 1 | Advances the output layer; this terminates the forward pass. |
| ToString | 1 | Returns a short description of the class labels held by this layer. |
03 Properties
| Name | Overloads | Summary |
|---|---|---|
| sortedClasses | 1 | Class labels sorted by descending probability after a call to Output.getDecision(). |
| probabilities | 1 | Probabilities that correspond one to one with Output.sortedClasses. |
| type | 1 | Gets the layer kind, always LayerTypes.Output. |
04 Members
Int32(), String())Creates the output layer.
| Name | Type | Description |
|---|---|---|
inputTensorDims | Int32() | The dimensions |
classes | String() | The class labels in the order used by the network's training set. |
Sorts the class probabilities in descending order and returns the most likely class label.
The label of the class with the highest predicted probability.
Always throws because the output layer is the end of the network and cannot feed another layer.
This method never returns.
Advances the output layer; this terminates the forward pass.
This method never returns.
Returns a short description of the class labels held by this layer.
A text that lists the class count and the first few class labels.
Class labels sorted by descending probability after a call to Output.getDecision().
Probabilities that correspond one to one with Output.sortedClasses.
Gets the layer kind, always LayerTypes.Output.