Class Model object for the prediction stage of neural network algorithm.
More...
◆ ModelImpl() [1/2]
Constructs model object for the prediction stage of neural network from the list of forward stages of the layers and the list of connections between the layers
- Parameters
-
[in] | forwardLayers | List of forward stages of the layers |
[in] | nextLayers | List of next layers for each layer with corresponding index |
[in] | storeWeightsInTable | Weights and biases storage status. True if weights and biases of all layers stored in one numeric table. False otherwise |
◆ ModelImpl() [2/2]
◆ getNextLayers()
Returns list of connections between layers
- Returns
- List of next layers for each layer with corresponding index
◆ getWeightsAndBiases()
const data_management::NumericTablePtr getWeightsAndBiases |
( |
| ) |
const |
|
inline |
Returns table containing all neural network weights and biases
- Returns
- Table containing all neural network weights and biases
◆ setWeightsAndBiases()
services::Status setWeightsAndBiases |
( |
const data_management::NumericTablePtr & |
weightsAndBiases | ) |
|
|
inline |
Sets table containing all neural network weights and biases
- Parameters
-
[in] | weightsAndBiases | Table containing all neural network weights and biases |
- Returns
- Status of computations
◆ _forwardLayers
neural_networks::ForwardLayersPtr _forwardLayers |
|
protected |
List of forward layers of the network
◆ _nextLayers
List of edges connecting the layers in the network
◆ _storeWeightsInTable
bool _storeWeightsInTable |
|
protected |
Flag. True if weights and biases of all the layers are stored in one numeric table
The documentation for this class was generated from the following file: